Codex CLI (openai/codex) | Context7

Codex CLI

https://github.com/openai/codex

Admin

OpenAI Codex CLI is a lightweight, local coding agent that runs in your terminal, enabling users to...

Tokens:431,483

Snippets:2,740

Trust Score:9.1

License:Apache-2.0

Update:21 hours ago

Show doc for...

Tokens:

Raw

### Run quickstart constructor example

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Execute the quickstart constructor example for an initial sanity check.

```bash
python examples/01_quickstart_constructor/sync.py
python examples/01_quickstart_constructor/async.py
```

--------------------------------

### Run example scripts

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Run synchronous or asynchronous versions of any example from the sdk/python directory.

```bash
python examples/<example-folder>/sync.py
python examples/<example-folder>/async.py
```

--------------------------------

### Build and install bubblewrap using meson

Source: https://github.com/openai/codex/blob/main/codex-rs/vendor/bubblewrap/README.md

Builds bubblewrap from source, runs test verification, and installs the resulting binaries into the system.

```sh
meson _builddir
meson compile -C _builddir
meson test -C _builddir
meson install -C _builddir
```

--------------------------------

### Install development dependencies from repository checkout

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Installs and activates development dependencies using uv in a local repository checkout.

```bash
cd sdk/python
uv sync --group dev
source .venv/bin/activate
```

--------------------------------

### Install development dependencies from checkout

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Sync development dependencies using uv and activate the virtual environment from sdk/python.

```bash
uv sync --group dev
source .venv/bin/activate
```

--------------------------------

### Define setup.json board template

Source: https://github.com/openai/codex/blob/main/codex-rs/ext/agent-message-board/README.md

Example structure of setup.json defining initial channels and descriptions. An optional top-level timestamp can define the creation time.

```json
{
  "template": {
    "version": 1,
    "channels": { "policy": { "description": "Shared policies" } }
  }
}
```

--------------------------------

### Start a process with process/start

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Spawns a process with specified arguments, working directory, and environment settings. Output from the process is streamed back via process/output notifications.

```json
{"id":2,"method":"process/start","params":{"processId":"proc-1","argv":["bash","-lc","printf 'ready\n'; while IFS= read -r line; do printf 'echo:%s\n' \"$line\"; done"],"cwd":"file:///tmp","env":{"PATH":"/usr/bin:/bin"},"tty":true,"pipeStdin":false,"arg0":null}}
{"id":2,"result":{"processId":"proc-1"}}
{"method":"process/output","params":{"processId":"proc-1","seq":1,"stream":"stdout","chunk":"cmVhZHkK"}}
```

--------------------------------

### Build and run codex app-server test client quickstart

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-test-client/README.md

Builds the debug codex binary, launches the websocket app-server in the background, and queries it with model-list. Execute these commands from the <reporoot>/codex-rs directory.

```bash
# 1) Build debug codex binary
cargo build -p codex-cli --bin codex

# 2) Start websocket app-server in background
cargo run -p codex-app-server-test-client -- \
  --codex-bin ./target/debug/codex \
  serve --listen ws://127.0.0.1:4222 --kill

# 3) Call app-server (defaults to ws://127.0.0.1:4222)
cargo run -p codex-app-server-test-client -- model-list
```

--------------------------------

### process/start

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Starts a new managed process with the specified command, environment variables, working directory, and terminal configurations.

```APIDOC
## RPC Method: `process/start`

### Description
Starts a new managed process within the connection lifecycle.

### Request Parameters
- **processId** (string) - Required. Caller-chosen stable id for this process within the connection. Reusing an existing ID is rejected.
- **argv** (array of strings) - Required. Non-empty command vector.
- **cwd** (string) - `file:` URI for the child process working directory.
- **env** (object) - Environment variables passed to the child process.
- **tty** (boolean) - When `true`, spawns a PTY-backed interactive process.
- **pipeStdin** (boolean) - When `true`, keeps non-PTY stdin writable via `process/write`.
- **arg0** (string | null) - Optional argv0 override forwarded to `codex-utils-pty`.

### Request Example
```json
{
  "processId": "proc-1",
  "argv": ["bash", "-lc", "printf 'hello\\n'"],
  "cwd": "file:///absolute/working/directory",
  "env": {
    "PATH": "/usr/bin:/bin"
  },
  "tty": true,
  "pipeStdin": false,
  "arg0": null
}
```

### Response
- **processId** (string) - The identifier of the started process.

### Response Example
```json
{
  "processId": "proc-1"
}
```
```

--------------------------------

### Run standalone attach example with cargo

Source: https://github.com/openai/codex/blob/main/codex-rs/cloud-client/README.md

Runs the attach example against an existing authorized thread and native gRPC origin. Requires CODEX_CLOUD_ACCESS_TOKEN and CODEX_CLOUD_ACCOUNT_ID to be set in your environment.

```sh
cargo run -p codex-cloud-client --example attach -- "$CODEX_CLOUD_GRPC_ENDPOINT" "$CODEX_CLOUD_THREAD_ID"
```

--------------------------------

### Install the published SDK

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Install the openai-codex package and its pinned runtime dependency using pip. Requires Python >=3.10.

```bash
python -m pip install openai-codex
```

--------------------------------

### Install optional Pillow package with uv

Source: https://github.com/openai/codex/blob/main/codex-rs/skills/src/assets/samples/imagegen/SKILL.md

Installs the optional `pillow` package used for image inspection and downscaling.

```bash
uv pip install pillow
```

--------------------------------

### Configure agent board with setup.json CLI command

Source: https://github.com/openai/codex/blob/main/codex-rs/ext/agent-message-board/README.md

Loads initial channels for an empty chat before starting agents. Use the same SQLite home directory as Codex (normally $CODEX_HOME or ~/.codex).

```sh
just agent-board-config ~/.codex <root-session-id> /absolute/path/setup.json
```

--------------------------------

### Install dependencies and build Codex CLI binary

Source: https://github.com/openai/codex/blob/main/scripts/mcp_conformance/README.md

Installs frozen workspace dependencies using pnpm and builds the locked Codex CLI binary using cargo.

```bash
pnpm install --frozen-lockfile
cargo build --locked --manifest-path codex-rs/Cargo.toml -p codex-cli --bin codex
```

--------------------------------

### Install Codex CLI via Homebrew

Source: https://github.com/openai/codex/blob/main/README.md

Installs the Codex desktop cask on macOS using Homebrew.

```shell
# Install using Homebrew
brew install --cask codex
```

--------------------------------

### Install openai-codex with pip

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Installs the OpenAI Codex Python SDK and its runtime dependencies. Requires Python >= 3.10.

```bash
pip install openai-codex
```

--------------------------------

### NetworkProxy::run() & Handle::shutdown()

Source: https://github.com/openai/codex/blob/main/codex-rs/network-proxy/README.md

Starts the network proxy listener and provides lifecycle control to gracefully terminate the proxy handle.

```APIDOC
## NetworkProxy Lifecycle

### Description
Starts the proxy service execution loop and shuts it down asynchronously.

### Methods
- **proxy.run() -> Result<Handle, Error>** - Spawns the proxy server and returns a control handle.
- **handle.shutdown() -> Result<(), Error>** - Initiates graceful shutdown of the running network proxy.

### Example
```rust
let handle = proxy.run().await?;
handle.shutdown().await?;
```
```

--------------------------------

### Install required OpenAI package with uv

Source: https://github.com/openai/codex/blob/main/codex-rs/skills/src/assets/samples/imagegen/SKILL.md

Installs the required `openai` Python package for CLI fallback operations.

```bash
uv pip install openai
```

--------------------------------

### Run turns asynchronously with AsyncCodex

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Demonstrates starting a thread and running a turn using the asynchronous AsyncCodex context manager.

```python
import asyncio

from openai_codex import AsyncCodex, Sandbox


async def main() -> None:
    async with AsyncCodex() as codex:
        thread = await codex.thread_start(sandbox=Sandbox.workspace_write)
        result = await thread.run("Continue where we left off.")
        print(result.final_response)


asyncio.run(main())
```

--------------------------------

### Manage async thread lifecycle, branching, and compaction

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Demonstrates how to start turns, resume, list, archive, unarchive, fork, and compact threads asynchronously using AsyncCodex.

```python
# Cell 9: full thread lifecycle and branching (async)
import asyncio


async def async_lifecycle_demo():
    async with AsyncCodex() as codex:
        thread = await codex.thread_start(model='gpt-5.4', config={'model_reasoning_effort': 'high'})
        first = await (await thread.turn('One sentence about structured planning.')).run()
        second = await (await thread.turn('Now restate it for a junior engineer.')).run()

        reopened = await codex.thread_resume(thread.id)
        listing_active = await codex.thread_list(limit=20, archived=False)
        reading = await reopened.read(include_turns=True)

        _ = await reopened.set_name('sdk-lifecycle-demo')
        _ = await codex.thread_archive(reopened.id)
        listing_archived = await codex.thread_list(limit=20, archived=True)
        unarchived = await codex.thread_unarchive(reopened.id)

        resumed = await codex.thread_resume(
            unarchived.id,
            model='gpt-5.4',
            config={'model_reasoning_effort': 'high'},
        )
        resumed_result = await (await resumed.turn('Continue in one short sentence.')).run()

        forked = await codex.thread_fork(unarchived.id, model='gpt-5.4')
        forked_result = await (await forked.turn('Take a different angle in one short sentence.')).run()

        compact_result = await unarchived.compact()

        print('Lifecycle OK:', thread.id)
        print('first:', first.id, first.status)
        print('second:', second.id, second.status)
        print('read.turns:', len(reading.thread.turns))
        print('list.active:', len(listing_active.data))
        print('list.archived:', len(listing_archived.data))
        print('resumed:', resumed_result.id, resumed_result.status)
        print('forked:', forked_result.id, forked_result.status)
        print('compact:', compact_result.model_dump(mode='json', by_alias=True))


await async_lifecycle_demo()

```

--------------------------------

### Define build options in meson_options.txt

Source: https://github.com/openai/codex/blob/main/codex-rs/vendor/bubblewrap/meson_options.txt

Configures project build-time options in Bubblewrap using Meson. Allows toggling shell completions, man pages, SELinux, test suites, setuid compatibility, and custom installation paths.

```meson
option(
  'bash_completion',
  type : 'feature',
  description : 'install bash completion script',
  value : 'enabled',
)
option(
  'bash_completion_dir',
  type : 'string',
  description : 'install bash completion script in this directory',
  value : '',
)
option(
  'bwrapdir',
  type : 'string',
  description : 'install bwrap in this directory [default: bindir, or libexecdir in subprojects]',
)
option(
  'build_rpath',
  type : 'string',
  description : 'set a RUNPATH or RPATH on the bwrap executable',
)
option(
  'install_rpath',
  type : 'string',
  description : 'set a RUNPATH or RPATH on the bwrap executable',
)
option(
  'man',
  type : 'feature',
  description : 'generate man pages',
  value : 'auto',
)
option(
  'program_prefix',
  type : 'string',
  description : 'Prepend string to bwrap executable name, for use with subprojects',
)
option(
  'python',
  type : 'string',
  description : 'Path to Python 3, or empty to use python3',
)
option(
  'support_setuid',
  type : 'boolean',
  description : 'Support setuid mode (deprecated)',
  value : false,
)
option(
  'require_userns',
  type : 'boolean',
  description : 'require user namespaces by default when installed setuid',
  value : false,
)
option(
  'selinux',
  type : 'feature',
  description : 'enable optional SELINUX support',
  value : 'auto',
)
option(
  'tests',
  type : 'boolean',
  description : 'build tests',
  value : true,
)
option(
  'zsh_completion',
  type : 'feature',
  description : 'install zsh completion script',
  value : 'enabled',
)
option(
  'zsh_completion_dir',
  type : 'string',
  description : 'install zsh completion script in this directory',
  value : '',
)
```

--------------------------------

### Start a thread and run a prompt with Codex

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/api-reference.md

Initialize a Codex context manager, start a thread specifying the model and reasoning effort configuration, and execute a prompt to obtain the final response.

```python
from openai_codex import Codex

with Codex() as codex:
    thread = codex.thread_start(model="gpt-5.4", config={"model_reasoning_effort": "high"})
    result = thread.run("Say hello in one sentence.")
    print(result.final_response)
```

--------------------------------

### Bootstrap local SDK environment and runtime package

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Locates the local SDK Python directory, verifies Python 3.10+ requirement, installs or verifies runtime dependencies, and sets up sys.path.

```python
# Cell 1: bootstrap local SDK imports + pinned runtime package
import os
import sys
from pathlib import Path

if sys.version_info < (3, 10):
    raise RuntimeError(
        f'Notebook requires Python 3.10+; current interpreter is {sys.version.split()[0]}.'
    )

def _is_sdk_python_dir(path: Path) -> bool:
    return (path / 'pyproject.toml').exists() and (path / 'src' / 'openai_codex').exists()


def _find_sdk_python_dir(start: Path) -> Path | None:
    checked = set()

    def _consider(candidate: Path) -> Path | None:
        resolved = candidate.resolve()
        if resolved in checked:
            return None
        checked.add(resolved)
        if _is_sdk_python_dir(resolved):
            return resolved
        return None

    for candidate in [start, *start.parents]:
        found = _consider(candidate)
        if found is not None:
            return found

    for candidate in [start / 'sdk' / 'python', *(parent / 'sdk' / 'python' for parent in start.parents)]:
        found = _consider(candidate)
        if found is not None:
            return found

    env_dir = os.environ.get('CODEX_PYTHON_SDK_DIR')
    if env_dir:
        found = _consider(Path(env_dir).expanduser())
        if found is not None:
            return found

    return None


repo_python_dir = _find_sdk_python_dir(Path.cwd())
if repo_python_dir is None:
    raise RuntimeError('Could not locate sdk/python. Set CODEX_PYTHON_SDK_DIR to your sdk/python path.')

repo_python_str = str(repo_python_dir)
if repo_python_str not in sys.path:
    sys.path.insert(0, repo_python_str)

from _runtime_setup import ensure_runtime_package_installed

runtime_version = ensure_runtime_package_installed(
    sys.executable,
    repo_python_dir,
)

src_dir = repo_python_dir / 'src'
examples_dir = repo_python_dir / 'examples'
src_str = str(src_dir)
examples_str = str(examples_dir)
if src_str not in sys.path:
    sys.path.insert(0, src_str)
if examples_str not in sys.path:
    sys.path.insert(0, examples_str)

# Force fresh imports after SDK upgrades in the same notebook kernel.
for module_name in list(sys.modules):
    if module_name == 'openai_codex' or module_name.startswith('openai_codex.'):
        sys.modules.pop(module_name, None)

print('Kernel:', sys.executable)
print('SDK source:', src_dir)
print('Runtime package:', runtime_version)
```

--------------------------------

### Build Codex from source and run tests

Source: https://github.com/openai/codex/blob/main/docs/install.md

Clone the repository, install the Rust toolchain and workspace helper tools, build Codex, and execute formatting and test commands with justfile helpers. Avoid `--all-features` during routine local runs to reduce build time and disk usage.

```bash
# Clone the repository and navigate to the root of the Cargo workspace.
git clone https://github.com/openai/codex.git
cd codex/codex-rs

# Install the Rust toolchain, if necessary.
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
source "$HOME/.cargo/env"
rustup component add rustfmt
rustup component add clippy
# Install helper tools used by the workspace justfile:
cargo install --locked just
# DotSlash fetches pinned development tools such as buildifier on first use.
cargo install --locked dotslash
# Install nextest for the `just test` helper.
cargo install --locked cargo-nextest

# Build Codex.
cargo build

# Launch the TUI with a sample prompt.
cargo run --bin codex -- "explain this codebase to me"

# After making changes, use the root justfile helpers (they default to codex-rs):
just fmt
just fix -p <crate-you-touched>

# Run the relevant tests (project-specific is fastest), for example:
just test -p codex-tui
# `just test` runs the test suite via nextest:
just test
# Avoid `--all-features` for routine local runs because it increases build
# time and `target/` disk usage by compiling additional feature combinations.
```

--------------------------------

### Bootstrap app-server daemon on Windows via PowerShell

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-daemon/README.md

Installs Codex and bootstraps the daemon on Windows. Run in a non-elevated PowerShell terminal whose host allows breakaway.

```powershell
irm https://chatgpt.com/codex/install.ps1 | iex
$codexHome = if ($env:CODEX_HOME) { $env:CODEX_HOME } else { Join-Path $HOME '.codex' }
& "$codexHome\packages\standalone\current\bin\codex.exe" app-server daemon bootstrap --remote-control
```

--------------------------------

### Directory Structure for an apply-patch Test Case

Source: https://github.com/openai/codex/blob/main/codex-rs/apply-patch/tests/fixtures/scenarios/README.md

Demonstrates the directory layout for a single test case that creates a new file. Organizes the starting file state in input/, the operation in patch.txt, and the target final state in expected/.

```text
001_add/
  input/
    foo.md
  expected/
    foo.md
    bar.md
  patch.txt
```

--------------------------------

### Codex.thread_start(*, approval_mode=ApprovalMode.auto_review, base_instructions=None, config=None, cwd=None, developer_instructions=None, ephemeral=None, model=None, model_provider=None, personality=None, sandbox: Sandbox | None = None) -> Thread

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/api-reference.md

Starts a new thread session with configurable instructions, approval modes, and sandboxing.

```APIDOC
## Codex.thread_start

### Description
Starts a new conversation thread.

### Signature
`thread_start(*, approval_mode=ApprovalMode.auto_review, base_instructions=None, config=None, cwd=None, developer_instructions=None, ephemeral=None, model=None, model_provider=None, personality=None, sandbox: Sandbox | None = None) -> Thread`

### Parameters
- **approval_mode** (optional) - Approval mode policy. Defaults to `ApprovalMode.auto_review`.
- **base_instructions** (optional) - Base instructions for the model.
- **config** (optional) - Thread-level configuration.
- **cwd** (optional) - Current working directory.
- **developer_instructions** (optional) - Developer instructions.
- **ephemeral** (optional) - Indicates whether the thread is ephemeral.
- **model** (optional) - Specific model identifier.
- **model_provider** (optional) - Provider identifier for the model.
- **personality** (optional) - Deprecated. Tone personality setting (e.g. `Personality.none`, `Personality.friendly`, `Personality.pragmatic`).
- **sandbox** (`Sandbox | None`, optional) - Sandbox environment definition. Defaults to `None`.

### Return Value
- Returns `Thread`
```

--------------------------------

### Start a thread and run a turn

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Starts a thread configured with workspace write access and executes a single turn, returning the final response and item count.

```python
from openai_codex import Codex, Sandbox

with Codex() as codex:
    thread = codex.thread_start(sandbox=Sandbox.workspace_write)
    result = thread.run("Say hello in one sentence.")

    print("Thread:", thread.id)
    print("Text:", result.final_response)
    print("Items:", len(result.items))
```

--------------------------------

### Preview diagram with render example

Source: https://github.com/openai/codex/blob/main/codex-rs/mermaid/README.md

Previews a Mermaid diagram file from codex-rs, optionally passing an available canvas width limit (such as 180).

```sh
cargo run -p codex-mermaid --example render -- 180 < diagram.mmd

```

--------------------------------

### Start a managed process via process/start

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Spawns a new managed child process using specified arguments, environment variables, working directory, and PTY or piped stdin settings. Reusing an existing processId is rejected.

```json
{
  "processId": "proc-1",
  "argv": ["bash", "-lc", "printf 'hello\\n'"],
  "cwd": "file:///absolute/working/directory",
  "env": {
    "PATH": "/usr/bin:/bin"
  },
  "tty": true,
  "pipeStdin": false,
  "arg0": null
}
```

```json
{
  "processId": "proc-1"
}
```

--------------------------------

### Run Standalone Network Proxy with JSON Configuration

Source: https://github.com/openai/codex/blob/main/codex-rs/network-proxy/README.md

Define a network policy in a standalone JSON file and start the proxy using the --config CLI option. The proxy requires network.enabled to be true and rejects unknown configuration fields.

```json
{
  "network": {
    "enabled": true,
    "proxy_url": "http://127.0.0.1:3128",
    "enable_socks5": false,
    "enable_socks5_udp": false,
    "allow_upstream_proxy": false,
    "allow_local_binding": false,
    "mode": "full",
    "mitm": false,
    "domains": {
      "api.example.com": "allow"
    }
  }
}
```

```bash
cargo run -p codex-network-proxy -- --config /path/to/network-proxy.json
```

--------------------------------

### Run a sandboxed shell with bwrap

Source: https://github.com/openai/codex/blob/main/codex-rs/vendor/bubblewrap/README.md

Creates a new mount namespace reusing the host's /usr directory with read-only access and starts a bash shell session in an isolated PID namespace.

```sh
bwrap \
    --ro-bind /usr /usr \
    --symlink usr/lib64 /lib64 \
    --proc /proc \
    --dev /dev \
    --unshare-pid \
    --new-session \
    bash
```

--------------------------------

### Run plugin analytics smoke test with plugin-analytics-smoke

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-test-client/README.md

Builds a debug binary and tests plugin enable/disable config writes, installation, and mentions while capturing analytics to a local JSONL file. Requires the plugin to be remotely installed and enabled with an authenticated profile.

```bash
# Build a debug Codex binary; analytics capture is unavailable in release builds.
cargo build -p codex-cli --bin codex

cargo run -p codex-app-server-test-client -- \
  --codex-bin ./target/debug/codex \
  plugin-analytics-smoke \
  --plugin-id linear@openai-curated-remote
```

--------------------------------

### Bootstrap app-server daemon on Linux or macOS

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-daemon/README.md

Installs Codex via curl and bootstraps the app-server daemon with remote control on a new Linux or macOS machine.

```sh
curl -fsSL https://chatgpt.com/codex/install.sh | sh
$HOME/.codex/packages/standalone/current/codex app-server daemon bootstrap --remote-control
```

--------------------------------

### Manage app-server daemon lifecycle with CLI commands

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-daemon/README.md

Execute daemon lifecycle commands including start, restart, update, remote control toggling, stop, version inspection, and bootstrap. Each successful command writes exactly one JSON object to stdout containing lifecycle and version metadata.

```sh
codex app-server daemon start
codex app-server daemon restart
codex app-server daemon update
codex app-server daemon enable-remote-control
codex app-server daemon disable-remote-control
codex app-server daemon stop
codex app-server daemon version
codex app-server daemon bootstrap --remote-control
```

--------------------------------

### Run codex-responses-api-proxy against Azure OpenAI

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Example for Azure deployments that accept Authorization: Bearer <key>. It unsets AZURE_OPENAI_API_KEY from the environment after piping it into the proxy, enables GET /shutdown, writes server info to /tmp/server-info.json, and forwards to the Azure responses endpoint.

```shell
printenv AZURE_OPENAI_API_KEY | env -u AZURE_OPENAI_API_KEY codex-responses-api-proxy \
  --http-shutdown \
  --server-info /tmp/server-info.json \
  --upstream-url "https://YOUR_PROJECT_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT/responses?api-version=2025-04-01-preview"
```

--------------------------------

### Start proxy as privileged user with --http-shutdown

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Run the proxy as a privileged user (root or sudo) with access to OPENAI_API_KEY. The proxy reads the auth token from stdin. The --http-shutdown flag allows an unprivileged user to shut down the server via GET /shutdown.

```shell
printenv OPENAI_API_KEY | env -u OPENAI_API_KEY codex-responses-api-proxy --http-shutdown --server-info /tmp/server-info.json
```

--------------------------------

### Install @openai/codex-sdk with npm

Source: https://github.com/openai/codex/blob/main/sdk/typescript/README.md

Installs the Codex TypeScript SDK package. Requires Node.js 18 or higher.

```bash
npm install @openai/codex-sdk
```

--------------------------------

### Build, sign, and run the native hardware test bundle

Source: https://github.com/openai/codex/blob/main/codex-rs/user-verification/README.md

Builds the hardware test binary, packages it into an app bundle with entitlements copied from TEMPLATE_APP, signs it with SIGNING_IDENTITY, and executes the ignored test.

```bash
set -euo pipefail
: "${TEMPLATE_APP:?Set the path to the isolated signed development app}"
: "${SIGNING_IDENTITY:?Set its Apple Development identity}"
run_dir=$(mktemp -d)
app="$run_dir/HardwareTest.app"
(cd codex-rs && cargo build --locked -p codex-user-verification --test hardware \
  --message-format=json) > "$run_dir/build.jsonl"
binary=$(jq -r 'select(.reason == "compiler-artifact" and .target.name == "hardware" and .executable != null) | .executable' "$run_dir/build.jsonl")
mkdir -p "$app/Contents/MacOS"
cp "$binary" "$app/Contents/MacOS/hardware"
cp "$TEMPLATE_APP/Contents/Info.plist" "$app/Contents/Info.plist"
cp "$TEMPLATE_APP/Contents/embedded.provisionprofile" "$app/Contents/embedded.provisionprofile"
/usr/libexec/PlistBuddy -c 'Set :CFBundleExecutable hardware' "$app/Contents/Info.plist"
codesign -d --xml --entitlements - "$TEMPLATE_APP" > "$run_dir/entitlements.plist"
codesign --force --options runtime --entitlements "$run_dir/entitlements.plist" \
  --sign "$SIGNING_IDENTITY" "$app"
codesign --verify --strict "$app"
echo "Test bundle: $app"
"$app/Contents/MacOS/hardware" --ignored --exact \
  persisted_key_signing_cancellation_and_retry --nocapture --test-threads=1
```

--------------------------------

### Launch proxy and configure Codex to use it

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Build and run the proxy, then add the provider configuration to ~/.codex/config.toml and use it with codex -p proxy.

```shell
# Launch the proxy, dump request/response pairs to /tmp/proxy
cd path/to/codex/codex-rs
cargo build
echo $OPENAI_API_KEY | ./target/debug/codex-responses-api-proxy \
    --port 60001 \
    --dump-dir /tmp/proxy


# Add this to ~/.codex/config.toml:

[model_providers.codex-responses-api-proxy]
name = 'codex-responses-api-proxy'
base_url = 'http://127.0.0.1:60001/v1'
wire_api='responses'

[profiles.proxy]
model_provider = "codex-responses-api-proxy"


# Use it
codex -p proxy
```

--------------------------------

### Install Codex CLI via npm

Source: https://github.com/openai/codex/blob/main/README.md

Installs Codex CLI globally using the npm package manager.

```shell
# Install using npm
npm install -g @openai/codex
```

--------------------------------

### Install Codex CLI on macOS or Linux

Source: https://github.com/openai/codex/blob/main/README.md

Downloads and runs the standalone install script for macOS or Linux environments.

```shell
curl -fsSL https://chatgpt.com/codex/install.sh | sh
```

--------------------------------

### Exercise generic remote configuration with run_bazel_with_buildbuddy.py

Source: https://github.com/openai/codex/blob/main/codex-rs/docs/bazel.md

Runs a Bazel build through the BuildBuddy wrapper using an API key and a repository override to invoke the generic remote host configuration.

```bash
BUILDBUDDY_API_KEY=... GITHUB_REPOSITORY=my-fork/codex \
  ./.github/scripts/run_bazel_with_buildbuddy.py \
  build --config=ci-linux //codex-rs/cli:codex
```

--------------------------------

### Inspect SDK documentation using help()

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Uses Python's built-in help function to inspect the SDK and core classes.

```python
import openai_codex
from openai_codex import Codex, CodexConfig

help(openai_codex)
help(Codex)
help(CodexConfig)
```

--------------------------------

### Install Codex CLI on Windows via PowerShell

Source: https://github.com/openai/codex/blob/main/README.md

Runs the Windows installation script directly in PowerShell with execution policy bypassed.

```shell
powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex"
```

--------------------------------

### Force GitHub Releases download on Windows PowerShell installer

Source: https://github.com/openai/codex/blob/main/README.md

Sets the environment variable to force GitHub Releases download before running the PowerShell installer.

```powershell
$env:CODEX_INSTALLER_USE_RELEASES_OPENAI_COM='false'; irm https://chatgpt.com/codex/install.ps1 | iex
```

--------------------------------

### Force GitHub Releases download on macOS/Linux installer

Source: https://github.com/openai/codex/blob/main/README.md

Installs Codex CLI on macOS or Linux while forcing downloads from GitHub Releases instead of releases.openai.com.

```shell
curl -fsSL https://chatgpt.com/codex/install.sh | CODEX_INSTALLER_USE_RELEASES_OPENAI_COM=false sh
```

--------------------------------

### Inspect SDK documentation using pydoc

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Runs pydoc CLI to display SDK module documentation.

```bash
python -m pydoc openai_codex
```

--------------------------------

### Initialize OtelProvider with OtelSettings and tracing_subscriber

Source: https://github.com/openai/codex/blob/main/codex-rs/otel/README.md

Constructs an OtelSettings instance and initializes an OtelProvider, attaching its logger and tracing layers to the tracing_subscriber registry. The provider also configures metrics when enabled.

```rust
use codex_otel::config::OtelExporter;
use codex_otel::config::OtelHttpProtocol;
use codex_otel::config::OtelSettings;
use codex_otel::OtelProvider;
use tracing_subscriber::prelude::*;

let settings = OtelSettings {
    environment: "dev".to_string(),
    service_name: "codex-cli".to_string(),
    service_version: env!("CARGO_PKG_VERSION").to_string(),
    codex_home: std::path::PathBuf::from("/tmp"),
    exporter: OtelExporter::OtlpHttp {
        endpoint: "https://otlp.example.com".to_string(),
        headers: std::collections::HashMap::new(),
        protocol: OtelHttpProtocol::Binary,
        tls: None,
    },
    trace_exporter: OtelExporter::OtlpHttp {
        endpoint: "https://otlp.example.com".to_string(),
        headers: std::collections::HashMap::new(),
        protocol: OtelHttpProtocol::Binary,
        tls: None,
    },
    metrics_exporter: OtelExporter::None,
    span_attributes: std::collections::BTreeMap::new(),
    tracestate: std::collections::BTreeMap::new(),
};

if let Some(provider) = OtelProvider::try_new(&settings)? {
    let registry = tracing_subscriber::registry()
        .with(provider.logger_layer())
        .with(provider.tracing_layer());
    registry.init();
}
```

--------------------------------

### Start a thread and run a turn with Codex

Source: https://github.com/openai/codex/blob/main/sdk/python/README.md

Reuses existing Codex authentication to start a thread and execute a turn. Returns a TurnResult containing the final response, collected items, and token usage.

```python
from openai_codex import Codex

with Codex() as codex:
    thread = codex.thread_start()
    result = thread.run("Explain this repository in three bullets.")
    print(result.final_response)
```

--------------------------------

### Test remote install and uninstall analytics with plugin-analytics-mutation-smoke

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-test-client/README.md

Performs a live smoke test by temporarily installing and uninstalling a remote plugin to validate mutation analytics events. Provide the backend identifier to `--remote-plugin-id` rather than the local plugin ID.

```bash
cargo run -p codex-app-server-test-client -- \
  --codex-bin ./target/debug/codex \
  plugin-analytics-mutation-smoke \
  --remote-plugin-id <REMOTE_PLUGIN_ID> \
  --confirm-account-mutation \
  --capture-file /tmp/plugin-mutation-analytics.jsonl
```

--------------------------------

### Using AsyncCodex Context Manager for Preferred Usage

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/api-reference.md

AsyncCodex initializes lazily; context entry is standard to ensure startup and shutdown are explicitly paired.

```python
async with AsyncCodex() as codex:
    ...
```

--------------------------------

### GET /shutdown

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Shuts down the proxy process with exit code 0. Only available when the `--http-shutdown` flag is set.

```APIDOC
## GET /shutdown

### Description
Exits the proxy process with code `0`. Only enabled when the `--http-shutdown` flag is provided.

### Method
GET

### Endpoint
/shutdown

### Parameters
None.

### Response
#### Success Response
Process exits with code `0`.

### Notes
- This endpoint is only available if `--http-shutdown` is set.
```

--------------------------------

### Execute synchronous single-turn conversation

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Starts a thread with specified model configuration and executes a single prompt using thread.run.

```python
# Cell 3: simple sync conversation
with Codex() as codex:
    thread = codex.thread_start(model='gpt-5.4', config={'model_reasoning_effort': 'high'})
    result = thread.run('Explain gradient descent in 3 bullets.')
    print(result.final_response)
```

--------------------------------

### GET /shutdown

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Terminates the proxy process with exit code 0. This endpoint is only available when the --http-shutdown option is enabled.

```APIDOC
## GET /shutdown

### Description
Terminates the proxy process with exit code 0. This endpoint is only available when the --http-shutdown option is enabled.

### Method
GET

### Endpoint
/shutdown

### Parameters

#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Request Example
GET /shutdown

### Response

#### Success Response
The process terminates with exit code 0.

#### Response Example
No response body.
```

--------------------------------

### Configuring BuildBuddy API key in ~/.bazelrc

Source: https://github.com/openai/codex/blob/main/codex-rs/docs/bazel.md

Configures the BuildBuddy API key in ~/.bazelrc or %workspace%/user.bazelrc for authentication. Keeping credentials in ~/.bazelrc prevents accidentally committing them to the workspace repository.

```bazelrc
# Local machine only; this file contains a BuildBuddy credential.
common --remote_header=x-buildbuddy-api-key=<your-buildbuddy-api-key>
```

--------------------------------

### account/gatewayOAuth/login

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server/README.md

Starts browser authorization for gateway OAuth for the current provider. Returns once credentials have been saved and emits status notifications during authorization.

```APIDOC
## account/gatewayOAuth/login

### Description
Starts the browser authorization flow for gateway OAuth. The initiating connection receives a `started` notification with `authUrl` to open in a browser. A second login request fails while a login is active.

### Type
RPC Method

### Parameters
None. Takes no params.

### Response
Returns `{}` after the credential has been saved.
```

--------------------------------

### Create Hero Image Generation Prompt

Source: https://github.com/openai/codex/blob/main/codex-rs/skills/src/assets/samples/imagegen/SKILL.md

Sample prompt applying the scaffolding schema to generate a clean product photograph of a ceramic coffee mug for a landing page.

```text
Use case: product-mockup
Asset type: landing page hero
Primary request: a minimal hero image of a ceramic coffee mug
Style/medium: clean product photography
Composition/framing: wide composition with usable negative space for page copy if needed
Lighting/mood: soft studio lighting
Constraints: no logos, no text, no watermark
```

--------------------------------

### Authenticate using browser login handle lifecycle

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Initiates ChatGPT login flow using login_chatgpt, waits for completion, and inspects account details.

```python
# Cell 2b: browser login handle lifecycle
with Codex() as codex:
    # Open this URL and call `wait()` without canceling when completing login for real.
    login = codex.login_chatgpt()
    print('Please complete login at:', login.auth_url)
    completed = login.wait()
    account = codex.account()

    print('login.id:', login.login_id)
    print('login.auth_url:', login.auth_url)
    print('login.completed.success:', completed.success)
    print('account:', account.email)
```

--------------------------------

### Configure HTTP and HTTPS Client Proxy Environment Variables

Source: https://github.com/openai/codex/blob/main/codex-rs/network-proxy/README.md

Direct HTTP, HTTPS, and WebSocket client traffic to the running proxy listener. Export these variables in your client environment.

```bash
export HTTP_PROXY="http://127.0.0.1:3128"
export HTTPS_PROXY="http://127.0.0.1:3128"
export WS_PROXY="http://127.0.0.1:3128"
export WSS_PROXY="http://127.0.0.1:3128"
```

--------------------------------

### Shutdown proxy via curl

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

When finished, an unprivileged user can shut down the proxy server using curl to GET /shutdown, since kill -SIGTERM is not available.

```shell
curl --fail --silent --show-error "${PROXY_BASE_URL}/shutdown"
```

--------------------------------

### Control streaming turns with steer and interrupt

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Shows how to guide an in-progress streaming turn using steer and cancel or halt turn generation using interrupt.

```python
# Cell 10: async turn controls (steer + interrupt)
import asyncio


async def async_stream_demo():
    async with AsyncCodex() as codex:
        thread = await codex.thread_start(model='gpt-5.4', config={'model_reasoning_effort': 'high'})
        steer_turn = await thread.turn('Count from 1 to 40 with commas, then one summary sentence.')

        steer_result = await steer_turn.steer('Keep it brief and stop after 10 numbers.')

        steer_event_count = 0
        steer_completed_status = None
        steer_deltas = []
        async for event in steer_turn.stream():
            steer_event_count += 1
            if event.method == 'item/agentMessage/delta':
                steer_deltas.append(event.payload.delta)
                continue
            if event.method == 'turn/completed':
                steer_completed_status = event.payload.turn.status.value

        if steer_completed_status is None:
            raise RuntimeError('stream ended without turn/completed')
        steer_preview = ''.join(steer_deltas).strip()

        interrupt_turn = await thread.turn('Count from 1 to 200 with commas, then one summary sentence.')
        interrupt_result = await interrupt_turn.interrupt()

        interrupt_event_count = 0
        interrupt_completed_status = None
        interrupt_deltas = []
        async for event in interrupt_turn.stream():
            interrupt_event_count += 1
            if event.method == 'item/agentMessage/delta':
                interrupt_deltas.append(event.payload.delta)
                continue
            if event.method == 'turn/completed':
                interrupt_completed_status = event.payload.turn.status.value

        if interrupt_completed_status is None:
            raise RuntimeError('stream ended without turn/completed')
        interrupt_preview = ''.join(interrupt_deltas).strip()

        print('steer.result:', steer_result.model_dump(mode='json', by_alias=True))
        print('steer.final.status:', steer_completed_status)
        print('steer.events.count:', steer_event_count)
        print('steer.assistant.preview:', steer_preview)
        print('interrupt.result:', interrupt_result.model_dump(mode='json', by_alias=True))
        print('interrupt.final.status:', interrupt_completed_status)
        print('interrupt.events.count:', interrupt_event_count)
        print('interrupt.assistant.preview:', interrupt_preview)


await async_stream_demo()

```

--------------------------------

### Parse and render templates using Template and render

Source: https://github.com/openai/codex/blob/main/codex-rs/utils/template/README.md

Demonstrates parsing a template with placeholder substitutions and literal braces, rendering it with key-value pairs, and using the one-shot render function.

```rust
use codex_utils_template::Template;
use codex_utils_template::render;

let template = Template::parse(
    "Hello, {{ name }}.\nLiteral braces: {{{{ and }}}}.\nMode: {{ mode }}",
)?;

let rendered = template.render([
    ("name", "Codex"),
    ("mode", "strict"),
])?;

assert_eq!(
    rendered,
    "Hello, Codex.\nLiteral braces: {{ and }}.\nMode: strict"
);

let one_shot = render("Hi {{ who }}!", [("who", "there")])?;
assert_eq!(one_shot, "Hi there!");
# Ok::<(), Box<dyn std::error::Error>>(())
```

--------------------------------

### Initialize exec server session

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Initializes a client session with the exec server. The server responds with session and environment information before the client confirms initialization.

```json
{"id":1,"method":"initialize","params":{"clientName":"example-client"}}
{"id":1,"result":{"sessionId":"00000000-0000-4000-8000-000000000001","environmentInfo":{"shell":{"name":"bash","path":"/bin/bash"},"cwd":"file:///tmp"}}}
{"method":"initialized","params":{}}
```

Tokens:

Raw

Tokens:

Raw

### Run quickstart constructor example

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Execute the quickstart constructor example for an initial sanity check.

```bash
python examples/01_quickstart_constructor/sync.py
python examples/01_quickstart_constructor/async.py
```

--------------------------------

### Run example scripts

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Run synchronous or asynchronous versions of any example from the sdk/python directory.

```bash
python examples/<example-folder>/sync.py
python examples/<example-folder>/async.py
```

--------------------------------

### Build and install bubblewrap using meson

Source: https://github.com/openai/codex/blob/main/codex-rs/vendor/bubblewrap/README.md

Builds bubblewrap from source, runs test verification, and installs the resulting binaries into the system.

```sh
meson _builddir
meson compile -C _builddir
meson test -C _builddir
meson install -C _builddir
```

--------------------------------

### Install development dependencies from repository checkout

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Installs and activates development dependencies using uv in a local repository checkout.

```bash
cd sdk/python
uv sync --group dev
source .venv/bin/activate
```

--------------------------------

### Install development dependencies from checkout

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Sync development dependencies using uv and activate the virtual environment from sdk/python.

```bash
uv sync --group dev
source .venv/bin/activate
```

--------------------------------

### Define setup.json board template

Source: https://github.com/openai/codex/blob/main/codex-rs/ext/agent-message-board/README.md

Example structure of setup.json defining initial channels and descriptions. An optional top-level timestamp can define the creation time.

```json
{
  "template": {
    "version": 1,
    "channels": { "policy": { "description": "Shared policies" } }
  }
}
```

--------------------------------

### Start a process with process/start

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Spawns a process with specified arguments, working directory, and environment settings. Output from the process is streamed back via process/output notifications.

```json
{"id":2,"method":"process/start","params":{"processId":"proc-1","argv":["bash","-lc","printf 'ready\n'; while IFS= read -r line; do printf 'echo:%s\n' \"$line\"; done"],"cwd":"file:///tmp","env":{"PATH":"/usr/bin:/bin"},"tty":true,"pipeStdin":false,"arg0":null}}
{"id":2,"result":{"processId":"proc-1"}}
{"method":"process/output","params":{"processId":"proc-1","seq":1,"stream":"stdout","chunk":"cmVhZHkK"}}
```

--------------------------------

### Build and run codex app-server test client quickstart

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-test-client/README.md

Builds the debug codex binary, launches the websocket app-server in the background, and queries it with model-list. Execute these commands from the <reporoot>/codex-rs directory.

```bash
# 1) Build debug codex binary
cargo build -p codex-cli --bin codex

# 2) Start websocket app-server in background
cargo run -p codex-app-server-test-client -- \
  --codex-bin ./target/debug/codex \
  serve --listen ws://127.0.0.1:4222 --kill

# 3) Call app-server (defaults to ws://127.0.0.1:4222)
cargo run -p codex-app-server-test-client -- model-list
```

--------------------------------

### process/start

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Starts a new managed process with the specified command, environment variables, working directory, and terminal configurations.

```APIDOC
## RPC Method: `process/start`

### Description
Starts a new managed process within the connection lifecycle.

### Request Parameters
- **processId** (string) - Required. Caller-chosen stable id for this process within the connection. Reusing an existing ID is rejected.
- **argv** (array of strings) - Required. Non-empty command vector.
- **cwd** (string) - `file:` URI for the child process working directory.
- **env** (object) - Environment variables passed to the child process.
- **tty** (boolean) - When `true`, spawns a PTY-backed interactive process.
- **pipeStdin** (boolean) - When `true`, keeps non-PTY stdin writable via `process/write`.
- **arg0** (string | null) - Optional argv0 override forwarded to `codex-utils-pty`.

### Request Example
```json
{
  "processId": "proc-1",
  "argv": ["bash", "-lc", "printf 'hello\\n'"],
  "cwd": "file:///absolute/working/directory",
  "env": {
    "PATH": "/usr/bin:/bin"
  },
  "tty": true,
  "pipeStdin": false,
  "arg0": null
}
```

### Response
- **processId** (string) - The identifier of the started process.

### Response Example
```json
{
  "processId": "proc-1"
}
```
```

--------------------------------

### Run standalone attach example with cargo

Source: https://github.com/openai/codex/blob/main/codex-rs/cloud-client/README.md

Runs the attach example against an existing authorized thread and native gRPC origin. Requires CODEX_CLOUD_ACCESS_TOKEN and CODEX_CLOUD_ACCOUNT_ID to be set in your environment.

```sh
cargo run -p codex-cloud-client --example attach -- "$CODEX_CLOUD_GRPC_ENDPOINT" "$CODEX_CLOUD_THREAD_ID"
```

--------------------------------

### Install the published SDK

Source: https://github.com/openai/codex/blob/main/sdk/python/examples/README.md

Install the openai-codex package and its pinned runtime dependency using pip. Requires Python >=3.10.

```bash
python -m pip install openai-codex
```

--------------------------------

### Install optional Pillow package with uv

Source: https://github.com/openai/codex/blob/main/codex-rs/skills/src/assets/samples/imagegen/SKILL.md

Installs the optional `pillow` package used for image inspection and downscaling.

```bash
uv pip install pillow
```

--------------------------------

### Configure agent board with setup.json CLI command

Source: https://github.com/openai/codex/blob/main/codex-rs/ext/agent-message-board/README.md

Loads initial channels for an empty chat before starting agents. Use the same SQLite home directory as Codex (normally $CODEX_HOME or ~/.codex).

```sh
just agent-board-config ~/.codex <root-session-id> /absolute/path/setup.json
```

--------------------------------

### Install dependencies and build Codex CLI binary

Source: https://github.com/openai/codex/blob/main/scripts/mcp_conformance/README.md

Installs frozen workspace dependencies using pnpm and builds the locked Codex CLI binary using cargo.

```bash
pnpm install --frozen-lockfile
cargo build --locked --manifest-path codex-rs/Cargo.toml -p codex-cli --bin codex
```

--------------------------------

### Install Codex CLI via Homebrew

Source: https://github.com/openai/codex/blob/main/README.md

Installs the Codex desktop cask on macOS using Homebrew.

```shell
# Install using Homebrew
brew install --cask codex
```

--------------------------------

### Install openai-codex with pip

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Installs the OpenAI Codex Python SDK and its runtime dependencies. Requires Python >= 3.10.

```bash
pip install openai-codex
```

--------------------------------

### NetworkProxy::run() & Handle::shutdown()

Source: https://github.com/openai/codex/blob/main/codex-rs/network-proxy/README.md

Starts the network proxy listener and provides lifecycle control to gracefully terminate the proxy handle.

```APIDOC
## NetworkProxy Lifecycle

### Description
Starts the proxy service execution loop and shuts it down asynchronously.

### Methods
- **proxy.run() -> Result<Handle, Error>** - Spawns the proxy server and returns a control handle.
- **handle.shutdown() -> Result<(), Error>** - Initiates graceful shutdown of the running network proxy.

### Example
```rust
let handle = proxy.run().await?;
handle.shutdown().await?;
```
```

--------------------------------

### Install required OpenAI package with uv

Source: https://github.com/openai/codex/blob/main/codex-rs/skills/src/assets/samples/imagegen/SKILL.md

Installs the required `openai` Python package for CLI fallback operations.

```bash
uv pip install openai
```

--------------------------------

### Run turns asynchronously with AsyncCodex

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Demonstrates starting a thread and running a turn using the asynchronous AsyncCodex context manager.

```python
import asyncio

from openai_codex import AsyncCodex, Sandbox


async def main() -> None:
    async with AsyncCodex() as codex:
        thread = await codex.thread_start(sandbox=Sandbox.workspace_write)
        result = await thread.run("Continue where we left off.")
        print(result.final_response)


asyncio.run(main())
```

--------------------------------

### Manage async thread lifecycle, branching, and compaction

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Demonstrates how to start turns, resume, list, archive, unarchive, fork, and compact threads asynchronously using AsyncCodex.

```python
# Cell 9: full thread lifecycle and branching (async)
import asyncio


async def async_lifecycle_demo():
    async with AsyncCodex() as codex:
        thread = await codex.thread_start(model='gpt-5.4', config={'model_reasoning_effort': 'high'})
        first = await (await thread.turn('One sentence about structured planning.')).run()
        second = await (await thread.turn('Now restate it for a junior engineer.')).run()

        reopened = await codex.thread_resume(thread.id)
        listing_active = await codex.thread_list(limit=20, archived=False)
        reading = await reopened.read(include_turns=True)

        _ = await reopened.set_name('sdk-lifecycle-demo')
        _ = await codex.thread_archive(reopened.id)
        listing_archived = await codex.thread_list(limit=20, archived=True)
        unarchived = await codex.thread_unarchive(reopened.id)

        resumed = await codex.thread_resume(
            unarchived.id,
            model='gpt-5.4',
            config={'model_reasoning_effort': 'high'},
        )
        resumed_result = await (await resumed.turn('Continue in one short sentence.')).run()

        forked = await codex.thread_fork(unarchived.id, model='gpt-5.4')
        forked_result = await (await forked.turn('Take a different angle in one short sentence.')).run()

        compact_result = await unarchived.compact()

        print('Lifecycle OK:', thread.id)
        print('first:', first.id, first.status)
        print('second:', second.id, second.status)
        print('read.turns:', len(reading.thread.turns))
        print('list.active:', len(listing_active.data))
        print('list.archived:', len(listing_archived.data))
        print('resumed:', resumed_result.id, resumed_result.status)
        print('forked:', forked_result.id, forked_result.status)
        print('compact:', compact_result.model_dump(mode='json', by_alias=True))


await async_lifecycle_demo()

```

--------------------------------

### Define build options in meson_options.txt

Source: https://github.com/openai/codex/blob/main/codex-rs/vendor/bubblewrap/meson_options.txt

Configures project build-time options in Bubblewrap using Meson. Allows toggling shell completions, man pages, SELinux, test suites, setuid compatibility, and custom installation paths.

```meson
option(
  'bash_completion',
  type : 'feature',
  description : 'install bash completion script',
  value : 'enabled',
)
option(
  'bash_completion_dir',
  type : 'string',
  description : 'install bash completion script in this directory',
  value : '',
)
option(
  'bwrapdir',
  type : 'string',
  description : 'install bwrap in this directory [default: bindir, or libexecdir in subprojects]',
)
option(
  'build_rpath',
  type : 'string',
  description : 'set a RUNPATH or RPATH on the bwrap executable',
)
option(
  'install_rpath',
  type : 'string',
  description : 'set a RUNPATH or RPATH on the bwrap executable',
)
option(
  'man',
  type : 'feature',
  description : 'generate man pages',
  value : 'auto',
)
option(
  'program_prefix',
  type : 'string',
  description : 'Prepend string to bwrap executable name, for use with subprojects',
)
option(
  'python',
  type : 'string',
  description : 'Path to Python 3, or empty to use python3',
)
option(
  'support_setuid',
  type : 'boolean',
  description : 'Support setuid mode (deprecated)',
  value : false,
)
option(
  'require_userns',
  type : 'boolean',
  description : 'require user namespaces by default when installed setuid',
  value : false,
)
option(
  'selinux',
  type : 'feature',
  description : 'enable optional SELINUX support',
  value : 'auto',
)
option(
  'tests',
  type : 'boolean',
  description : 'build tests',
  value : true,
)
option(
  'zsh_completion',
  type : 'feature',
  description : 'install zsh completion script',
  value : 'enabled',
)
option(
  'zsh_completion_dir',
  type : 'string',
  description : 'install zsh completion script in this directory',
  value : '',
)
```

--------------------------------

### Start a thread and run a prompt with Codex

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/api-reference.md

Initialize a Codex context manager, start a thread specifying the model and reasoning effort configuration, and execute a prompt to obtain the final response.

```python
from openai_codex import Codex

with Codex() as codex:
    thread = codex.thread_start(model="gpt-5.4", config={"model_reasoning_effort": "high"})
    result = thread.run("Say hello in one sentence.")
    print(result.final_response)
```

--------------------------------

### Bootstrap local SDK environment and runtime package

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Locates the local SDK Python directory, verifies Python 3.10+ requirement, installs or verifies runtime dependencies, and sets up sys.path.

```python
# Cell 1: bootstrap local SDK imports + pinned runtime package
import os
import sys
from pathlib import Path

if sys.version_info < (3, 10):
    raise RuntimeError(
        f'Notebook requires Python 3.10+; current interpreter is {sys.version.split()[0]}.'
    )

def _is_sdk_python_dir(path: Path) -> bool:
    return (path / 'pyproject.toml').exists() and (path / 'src' / 'openai_codex').exists()


def _find_sdk_python_dir(start: Path) -> Path | None:
    checked = set()

    def _consider(candidate: Path) -> Path | None:
        resolved = candidate.resolve()
        if resolved in checked:
            return None
        checked.add(resolved)
        if _is_sdk_python_dir(resolved):
            return resolved
        return None

    for candidate in [start, *start.parents]:
        found = _consider(candidate)
        if found is not None:
            return found

    for candidate in [start / 'sdk' / 'python', *(parent / 'sdk' / 'python' for parent in start.parents)]:
        found = _consider(candidate)
        if found is not None:
            return found

    env_dir = os.environ.get('CODEX_PYTHON_SDK_DIR')
    if env_dir:
        found = _consider(Path(env_dir).expanduser())
        if found is not None:
            return found

    return None


repo_python_dir = _find_sdk_python_dir(Path.cwd())
if repo_python_dir is None:
    raise RuntimeError('Could not locate sdk/python. Set CODEX_PYTHON_SDK_DIR to your sdk/python path.')

repo_python_str = str(repo_python_dir)
if repo_python_str not in sys.path:
    sys.path.insert(0, repo_python_str)

from _runtime_setup import ensure_runtime_package_installed

runtime_version = ensure_runtime_package_installed(
    sys.executable,
    repo_python_dir,
)

src_dir = repo_python_dir / 'src'
examples_dir = repo_python_dir / 'examples'
src_str = str(src_dir)
examples_str = str(examples_dir)
if src_str not in sys.path:
    sys.path.insert(0, src_str)
if examples_str not in sys.path:
    sys.path.insert(0, examples_str)

# Force fresh imports after SDK upgrades in the same notebook kernel.
for module_name in list(sys.modules):
    if module_name == 'openai_codex' or module_name.startswith('openai_codex.'):
        sys.modules.pop(module_name, None)

print('Kernel:', sys.executable)
print('SDK source:', src_dir)
print('Runtime package:', runtime_version)
```

--------------------------------

### Build Codex from source and run tests

Source: https://github.com/openai/codex/blob/main/docs/install.md

Clone the repository, install the Rust toolchain and workspace helper tools, build Codex, and execute formatting and test commands with justfile helpers. Avoid `--all-features` during routine local runs to reduce build time and disk usage.

```bash
# Clone the repository and navigate to the root of the Cargo workspace.
git clone https://github.com/openai/codex.git
cd codex/codex-rs

# Install the Rust toolchain, if necessary.
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
source "$HOME/.cargo/env"
rustup component add rustfmt
rustup component add clippy
# Install helper tools used by the workspace justfile:
cargo install --locked just
# DotSlash fetches pinned development tools such as buildifier on first use.
cargo install --locked dotslash
# Install nextest for the `just test` helper.
cargo install --locked cargo-nextest

# Build Codex.
cargo build

# Launch the TUI with a sample prompt.
cargo run --bin codex -- "explain this codebase to me"

# After making changes, use the root justfile helpers (they default to codex-rs):
just fmt
just fix -p <crate-you-touched>

# Run the relevant tests (project-specific is fastest), for example:
just test -p codex-tui
# `just test` runs the test suite via nextest:
just test
# Avoid `--all-features` for routine local runs because it increases build
# time and `target/` disk usage by compiling additional feature combinations.
```

--------------------------------

### Bootstrap app-server daemon on Windows via PowerShell

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-daemon/README.md

Installs Codex and bootstraps the daemon on Windows. Run in a non-elevated PowerShell terminal whose host allows breakaway.

```powershell
irm https://chatgpt.com/codex/install.ps1 | iex
$codexHome = if ($env:CODEX_HOME) { $env:CODEX_HOME } else { Join-Path $HOME '.codex' }
& "$codexHome\packages\standalone\current\bin\codex.exe" app-server daemon bootstrap --remote-control
```

--------------------------------

### Directory Structure for an apply-patch Test Case

Source: https://github.com/openai/codex/blob/main/codex-rs/apply-patch/tests/fixtures/scenarios/README.md

Demonstrates the directory layout for a single test case that creates a new file. Organizes the starting file state in input/, the operation in patch.txt, and the target final state in expected/.

```text
001_add/
  input/
    foo.md
  expected/
    foo.md
    bar.md
  patch.txt
```

--------------------------------

### Codex.thread_start(*, approval_mode=ApprovalMode.auto_review, base_instructions=None, config=None, cwd=None, developer_instructions=None, ephemeral=None, model=None, model_provider=None, personality=None, sandbox: Sandbox | None = None) -> Thread

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/api-reference.md

Starts a new thread session with configurable instructions, approval modes, and sandboxing.

```APIDOC
## Codex.thread_start

### Description
Starts a new conversation thread.

### Signature
`thread_start(*, approval_mode=ApprovalMode.auto_review, base_instructions=None, config=None, cwd=None, developer_instructions=None, ephemeral=None, model=None, model_provider=None, personality=None, sandbox: Sandbox | None = None) -> Thread`

### Parameters
- **approval_mode** (optional) - Approval mode policy. Defaults to `ApprovalMode.auto_review`.
- **base_instructions** (optional) - Base instructions for the model.
- **config** (optional) - Thread-level configuration.
- **cwd** (optional) - Current working directory.
- **developer_instructions** (optional) - Developer instructions.
- **ephemeral** (optional) - Indicates whether the thread is ephemeral.
- **model** (optional) - Specific model identifier.
- **model_provider** (optional) - Provider identifier for the model.
- **personality** (optional) - Deprecated. Tone personality setting (e.g. `Personality.none`, `Personality.friendly`, `Personality.pragmatic`).
- **sandbox** (`Sandbox | None`, optional) - Sandbox environment definition. Defaults to `None`.

### Return Value
- Returns `Thread`
```

--------------------------------

### Start a thread and run a turn

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Starts a thread configured with workspace write access and executes a single turn, returning the final response and item count.

```python
from openai_codex import Codex, Sandbox

with Codex() as codex:
    thread = codex.thread_start(sandbox=Sandbox.workspace_write)
    result = thread.run("Say hello in one sentence.")

    print("Thread:", thread.id)
    print("Text:", result.final_response)
    print("Items:", len(result.items))
```

--------------------------------

### Preview diagram with render example

Source: https://github.com/openai/codex/blob/main/codex-rs/mermaid/README.md

Previews a Mermaid diagram file from codex-rs, optionally passing an available canvas width limit (such as 180).

```sh
cargo run -p codex-mermaid --example render -- 180 < diagram.mmd

```

--------------------------------

### Start a managed process via process/start

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Spawns a new managed child process using specified arguments, environment variables, working directory, and PTY or piped stdin settings. Reusing an existing processId is rejected.

```json
{
  "processId": "proc-1",
  "argv": ["bash", "-lc", "printf 'hello\\n'"],
  "cwd": "file:///absolute/working/directory",
  "env": {
    "PATH": "/usr/bin:/bin"
  },
  "tty": true,
  "pipeStdin": false,
  "arg0": null
}
```

```json
{
  "processId": "proc-1"
}
```

--------------------------------

### Run Standalone Network Proxy with JSON Configuration

Source: https://github.com/openai/codex/blob/main/codex-rs/network-proxy/README.md

Define a network policy in a standalone JSON file and start the proxy using the --config CLI option. The proxy requires network.enabled to be true and rejects unknown configuration fields.

```json
{
  "network": {
    "enabled": true,
    "proxy_url": "http://127.0.0.1:3128",
    "enable_socks5": false,
    "enable_socks5_udp": false,
    "allow_upstream_proxy": false,
    "allow_local_binding": false,
    "mode": "full",
    "mitm": false,
    "domains": {
      "api.example.com": "allow"
    }
  }
}
```

```bash
cargo run -p codex-network-proxy -- --config /path/to/network-proxy.json
```

--------------------------------

### Run a sandboxed shell with bwrap

Source: https://github.com/openai/codex/blob/main/codex-rs/vendor/bubblewrap/README.md

Creates a new mount namespace reusing the host's /usr directory with read-only access and starts a bash shell session in an isolated PID namespace.

```sh
bwrap \
    --ro-bind /usr /usr \
    --symlink usr/lib64 /lib64 \
    --proc /proc \
    --dev /dev \
    --unshare-pid \
    --new-session \
    bash
```

--------------------------------

### Run plugin analytics smoke test with plugin-analytics-smoke

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-test-client/README.md

Builds a debug binary and tests plugin enable/disable config writes, installation, and mentions while capturing analytics to a local JSONL file. Requires the plugin to be remotely installed and enabled with an authenticated profile.

```bash
# Build a debug Codex binary; analytics capture is unavailable in release builds.
cargo build -p codex-cli --bin codex

cargo run -p codex-app-server-test-client -- \
  --codex-bin ./target/debug/codex \
  plugin-analytics-smoke \
  --plugin-id linear@openai-curated-remote
```

--------------------------------

### Bootstrap app-server daemon on Linux or macOS

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-daemon/README.md

Installs Codex via curl and bootstraps the app-server daemon with remote control on a new Linux or macOS machine.

```sh
curl -fsSL https://chatgpt.com/codex/install.sh | sh
$HOME/.codex/packages/standalone/current/codex app-server daemon bootstrap --remote-control
```

--------------------------------

### Manage app-server daemon lifecycle with CLI commands

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-daemon/README.md

Execute daemon lifecycle commands including start, restart, update, remote control toggling, stop, version inspection, and bootstrap. Each successful command writes exactly one JSON object to stdout containing lifecycle and version metadata.

```sh
codex app-server daemon start
codex app-server daemon restart
codex app-server daemon update
codex app-server daemon enable-remote-control
codex app-server daemon disable-remote-control
codex app-server daemon stop
codex app-server daemon version
codex app-server daemon bootstrap --remote-control
```

--------------------------------

### Run codex-responses-api-proxy against Azure OpenAI

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Example for Azure deployments that accept Authorization: Bearer <key>. It unsets AZURE_OPENAI_API_KEY from the environment after piping it into the proxy, enables GET /shutdown, writes server info to /tmp/server-info.json, and forwards to the Azure responses endpoint.

```shell
printenv AZURE_OPENAI_API_KEY | env -u AZURE_OPENAI_API_KEY codex-responses-api-proxy \
  --http-shutdown \
  --server-info /tmp/server-info.json \
  --upstream-url "https://YOUR_PROJECT_NAME.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT/responses?api-version=2025-04-01-preview"
```

--------------------------------

### Start proxy as privileged user with --http-shutdown

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Run the proxy as a privileged user (root or sudo) with access to OPENAI_API_KEY. The proxy reads the auth token from stdin. The --http-shutdown flag allows an unprivileged user to shut down the server via GET /shutdown.

```shell
printenv OPENAI_API_KEY | env -u OPENAI_API_KEY codex-responses-api-proxy --http-shutdown --server-info /tmp/server-info.json
```

--------------------------------

### Install @openai/codex-sdk with npm

Source: https://github.com/openai/codex/blob/main/sdk/typescript/README.md

Installs the Codex TypeScript SDK package. Requires Node.js 18 or higher.

```bash
npm install @openai/codex-sdk
```

--------------------------------

### Build, sign, and run the native hardware test bundle

Source: https://github.com/openai/codex/blob/main/codex-rs/user-verification/README.md

Builds the hardware test binary, packages it into an app bundle with entitlements copied from TEMPLATE_APP, signs it with SIGNING_IDENTITY, and executes the ignored test.

```bash
set -euo pipefail
: "${TEMPLATE_APP:?Set the path to the isolated signed development app}"
: "${SIGNING_IDENTITY:?Set its Apple Development identity}"
run_dir=$(mktemp -d)
app="$run_dir/HardwareTest.app"
(cd codex-rs && cargo build --locked -p codex-user-verification --test hardware \
  --message-format=json) > "$run_dir/build.jsonl"
binary=$(jq -r 'select(.reason == "compiler-artifact" and .target.name == "hardware" and .executable != null) | .executable' "$run_dir/build.jsonl")
mkdir -p "$app/Contents/MacOS"
cp "$binary" "$app/Contents/MacOS/hardware"
cp "$TEMPLATE_APP/Contents/Info.plist" "$app/Contents/Info.plist"
cp "$TEMPLATE_APP/Contents/embedded.provisionprofile" "$app/Contents/embedded.provisionprofile"
/usr/libexec/PlistBuddy -c 'Set :CFBundleExecutable hardware' "$app/Contents/Info.plist"
codesign -d --xml --entitlements - "$TEMPLATE_APP" > "$run_dir/entitlements.plist"
codesign --force --options runtime --entitlements "$run_dir/entitlements.plist" \
  --sign "$SIGNING_IDENTITY" "$app"
codesign --verify --strict "$app"
echo "Test bundle: $app"
"$app/Contents/MacOS/hardware" --ignored --exact \
  persisted_key_signing_cancellation_and_retry --nocapture --test-threads=1
```

--------------------------------

### Launch proxy and configure Codex to use it

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Build and run the proxy, then add the provider configuration to ~/.codex/config.toml and use it with codex -p proxy.

```shell
# Launch the proxy, dump request/response pairs to /tmp/proxy
cd path/to/codex/codex-rs
cargo build
echo $OPENAI_API_KEY | ./target/debug/codex-responses-api-proxy \
    --port 60001 \
    --dump-dir /tmp/proxy


# Add this to ~/.codex/config.toml:

[model_providers.codex-responses-api-proxy]
name = 'codex-responses-api-proxy'
base_url = 'http://127.0.0.1:60001/v1'
wire_api='responses'

[profiles.proxy]
model_provider = "codex-responses-api-proxy"


# Use it
codex -p proxy
```

--------------------------------

### Install Codex CLI via npm

Source: https://github.com/openai/codex/blob/main/README.md

Installs Codex CLI globally using the npm package manager.

```shell
# Install using npm
npm install -g @openai/codex
```

--------------------------------

### Install Codex CLI on macOS or Linux

Source: https://github.com/openai/codex/blob/main/README.md

Downloads and runs the standalone install script for macOS or Linux environments.

```shell
curl -fsSL https://chatgpt.com/codex/install.sh | sh
```

--------------------------------

### Exercise generic remote configuration with run_bazel_with_buildbuddy.py

Source: https://github.com/openai/codex/blob/main/codex-rs/docs/bazel.md

Runs a Bazel build through the BuildBuddy wrapper using an API key and a repository override to invoke the generic remote host configuration.

```bash
BUILDBUDDY_API_KEY=... GITHUB_REPOSITORY=my-fork/codex \
  ./.github/scripts/run_bazel_with_buildbuddy.py \
  build --config=ci-linux //codex-rs/cli:codex
```

--------------------------------

### Inspect SDK documentation using help()

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Uses Python's built-in help function to inspect the SDK and core classes.

```python
import openai_codex
from openai_codex import Codex, CodexConfig

help(openai_codex)
help(Codex)
help(CodexConfig)
```

--------------------------------

### Install Codex CLI on Windows via PowerShell

Source: https://github.com/openai/codex/blob/main/README.md

Runs the Windows installation script directly in PowerShell with execution policy bypassed.

```shell
powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex"
```

--------------------------------

### Force GitHub Releases download on Windows PowerShell installer

Source: https://github.com/openai/codex/blob/main/README.md

Sets the environment variable to force GitHub Releases download before running the PowerShell installer.

```powershell
$env:CODEX_INSTALLER_USE_RELEASES_OPENAI_COM='false'; irm https://chatgpt.com/codex/install.ps1 | iex
```

--------------------------------

### Force GitHub Releases download on macOS/Linux installer

Source: https://github.com/openai/codex/blob/main/README.md

Installs Codex CLI on macOS or Linux while forcing downloads from GitHub Releases instead of releases.openai.com.

```shell
curl -fsSL https://chatgpt.com/codex/install.sh | CODEX_INSTALLER_USE_RELEASES_OPENAI_COM=false sh
```

--------------------------------

### Inspect SDK documentation using pydoc

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/getting-started.md

Runs pydoc CLI to display SDK module documentation.

```bash
python -m pydoc openai_codex
```

--------------------------------

### Initialize OtelProvider with OtelSettings and tracing_subscriber

Source: https://github.com/openai/codex/blob/main/codex-rs/otel/README.md

Constructs an OtelSettings instance and initializes an OtelProvider, attaching its logger and tracing layers to the tracing_subscriber registry. The provider also configures metrics when enabled.

```rust
use codex_otel::config::OtelExporter;
use codex_otel::config::OtelHttpProtocol;
use codex_otel::config::OtelSettings;
use codex_otel::OtelProvider;
use tracing_subscriber::prelude::*;

let settings = OtelSettings {
    environment: "dev".to_string(),
    service_name: "codex-cli".to_string(),
    service_version: env!("CARGO_PKG_VERSION").to_string(),
    codex_home: std::path::PathBuf::from("/tmp"),
    exporter: OtelExporter::OtlpHttp {
        endpoint: "https://otlp.example.com".to_string(),
        headers: std::collections::HashMap::new(),
        protocol: OtelHttpProtocol::Binary,
        tls: None,
    },
    trace_exporter: OtelExporter::OtlpHttp {
        endpoint: "https://otlp.example.com".to_string(),
        headers: std::collections::HashMap::new(),
        protocol: OtelHttpProtocol::Binary,
        tls: None,
    },
    metrics_exporter: OtelExporter::None,
    span_attributes: std::collections::BTreeMap::new(),
    tracestate: std::collections::BTreeMap::new(),
};

if let Some(provider) = OtelProvider::try_new(&settings)? {
    let registry = tracing_subscriber::registry()
        .with(provider.logger_layer())
        .with(provider.tracing_layer());
    registry.init();
}
```

--------------------------------

### Start a thread and run a turn with Codex

Source: https://github.com/openai/codex/blob/main/sdk/python/README.md

Reuses existing Codex authentication to start a thread and execute a turn. Returns a TurnResult containing the final response, collected items, and token usage.

```python
from openai_codex import Codex

with Codex() as codex:
    thread = codex.thread_start()
    result = thread.run("Explain this repository in three bullets.")
    print(result.final_response)
```

--------------------------------

### Test remote install and uninstall analytics with plugin-analytics-mutation-smoke

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server-test-client/README.md

Performs a live smoke test by temporarily installing and uninstalling a remote plugin to validate mutation analytics events. Provide the backend identifier to `--remote-plugin-id` rather than the local plugin ID.

```bash
cargo run -p codex-app-server-test-client -- \
  --codex-bin ./target/debug/codex \
  plugin-analytics-mutation-smoke \
  --remote-plugin-id <REMOTE_PLUGIN_ID> \
  --confirm-account-mutation \
  --capture-file /tmp/plugin-mutation-analytics.jsonl
```

--------------------------------

### Using AsyncCodex Context Manager for Preferred Usage

Source: https://github.com/openai/codex/blob/main/sdk/python/docs/api-reference.md

AsyncCodex initializes lazily; context entry is standard to ensure startup and shutdown are explicitly paired.

```python
async with AsyncCodex() as codex:
    ...
```

--------------------------------

### GET /shutdown

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Shuts down the proxy process with exit code 0. Only available when the `--http-shutdown` flag is set.

```APIDOC
## GET /shutdown

### Description
Exits the proxy process with code `0`. Only enabled when the `--http-shutdown` flag is provided.

### Method
GET

### Endpoint
/shutdown

### Parameters
None.

### Response
#### Success Response
Process exits with code `0`.

### Notes
- This endpoint is only available if `--http-shutdown` is set.
```

--------------------------------

### Execute synchronous single-turn conversation

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Starts a thread with specified model configuration and executes a single prompt using thread.run.

```python
# Cell 3: simple sync conversation
with Codex() as codex:
    thread = codex.thread_start(model='gpt-5.4', config={'model_reasoning_effort': 'high'})
    result = thread.run('Explain gradient descent in 3 bullets.')
    print(result.final_response)
```

--------------------------------

### GET /shutdown

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

Terminates the proxy process with exit code 0. This endpoint is only available when the --http-shutdown option is enabled.

```APIDOC
## GET /shutdown

### Description
Terminates the proxy process with exit code 0. This endpoint is only available when the --http-shutdown option is enabled.

### Method
GET

### Endpoint
/shutdown

### Parameters

#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Request Example
GET /shutdown

### Response

#### Success Response
The process terminates with exit code 0.

#### Response Example
No response body.
```

--------------------------------

### Configuring BuildBuddy API key in ~/.bazelrc

Source: https://github.com/openai/codex/blob/main/codex-rs/docs/bazel.md

Configures the BuildBuddy API key in ~/.bazelrc or %workspace%/user.bazelrc for authentication. Keeping credentials in ~/.bazelrc prevents accidentally committing them to the workspace repository.

```bazelrc
# Local machine only; this file contains a BuildBuddy credential.
common --remote_header=x-buildbuddy-api-key=<your-buildbuddy-api-key>
```

--------------------------------

### account/gatewayOAuth/login

Source: https://github.com/openai/codex/blob/main/codex-rs/app-server/README.md

Starts browser authorization for gateway OAuth for the current provider. Returns once credentials have been saved and emits status notifications during authorization.

```APIDOC
## account/gatewayOAuth/login

### Description
Starts the browser authorization flow for gateway OAuth. The initiating connection receives a `started` notification with `authUrl` to open in a browser. A second login request fails while a login is active.

### Type
RPC Method

### Parameters
None. Takes no params.

### Response
Returns `{}` after the credential has been saved.
```

--------------------------------

### Create Hero Image Generation Prompt

Source: https://github.com/openai/codex/blob/main/codex-rs/skills/src/assets/samples/imagegen/SKILL.md

Sample prompt applying the scaffolding schema to generate a clean product photograph of a ceramic coffee mug for a landing page.

```text
Use case: product-mockup
Asset type: landing page hero
Primary request: a minimal hero image of a ceramic coffee mug
Style/medium: clean product photography
Composition/framing: wide composition with usable negative space for page copy if needed
Lighting/mood: soft studio lighting
Constraints: no logos, no text, no watermark
```

--------------------------------

### Authenticate using browser login handle lifecycle

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Initiates ChatGPT login flow using login_chatgpt, waits for completion, and inspects account details.

```python
# Cell 2b: browser login handle lifecycle
with Codex() as codex:
    # Open this URL and call `wait()` without canceling when completing login for real.
    login = codex.login_chatgpt()
    print('Please complete login at:', login.auth_url)
    completed = login.wait()
    account = codex.account()

    print('login.id:', login.login_id)
    print('login.auth_url:', login.auth_url)
    print('login.completed.success:', completed.success)
    print('account:', account.email)
```

--------------------------------

### Configure HTTP and HTTPS Client Proxy Environment Variables

Source: https://github.com/openai/codex/blob/main/codex-rs/network-proxy/README.md

Direct HTTP, HTTPS, and WebSocket client traffic to the running proxy listener. Export these variables in your client environment.

```bash
export HTTP_PROXY="http://127.0.0.1:3128"
export HTTPS_PROXY="http://127.0.0.1:3128"
export WS_PROXY="http://127.0.0.1:3128"
export WSS_PROXY="http://127.0.0.1:3128"
```

--------------------------------

### Shutdown proxy via curl

Source: https://github.com/openai/codex/blob/main/codex-rs/responses-api-proxy/README.md

When finished, an unprivileged user can shut down the proxy server using curl to GET /shutdown, since kill -SIGTERM is not available.

```shell
curl --fail --silent --show-error "${PROXY_BASE_URL}/shutdown"
```

--------------------------------

### Control streaming turns with steer and interrupt

Source: https://github.com/openai/codex/blob/main/sdk/python/notebooks/sdk_walkthrough.ipynb

Shows how to guide an in-progress streaming turn using steer and cancel or halt turn generation using interrupt.

```python
# Cell 10: async turn controls (steer + interrupt)
import asyncio


async def async_stream_demo():
    async with AsyncCodex() as codex:
        thread = await codex.thread_start(model='gpt-5.4', config={'model_reasoning_effort': 'high'})
        steer_turn = await thread.turn('Count from 1 to 40 with commas, then one summary sentence.')

        steer_result = await steer_turn.steer('Keep it brief and stop after 10 numbers.')

        steer_event_count = 0
        steer_completed_status = None
        steer_deltas = []
        async for event in steer_turn.stream():
            steer_event_count += 1
            if event.method == 'item/agentMessage/delta':
                steer_deltas.append(event.payload.delta)
                continue
            if event.method == 'turn/completed':
                steer_completed_status = event.payload.turn.status.value

        if steer_completed_status is None:
            raise RuntimeError('stream ended without turn/completed')
        steer_preview = ''.join(steer_deltas).strip()

        interrupt_turn = await thread.turn('Count from 1 to 200 with commas, then one summary sentence.')
        interrupt_result = await interrupt_turn.interrupt()

        interrupt_event_count = 0
        interrupt_completed_status = None
        interrupt_deltas = []
        async for event in interrupt_turn.stream():
            interrupt_event_count += 1
            if event.method == 'item/agentMessage/delta':
                interrupt_deltas.append(event.payload.delta)
                continue
            if event.method == 'turn/completed':
                interrupt_completed_status = event.payload.turn.status.value

        if interrupt_completed_status is None:
            raise RuntimeError('stream ended without turn/completed')
        interrupt_preview = ''.join(interrupt_deltas).strip()

        print('steer.result:', steer_result.model_dump(mode='json', by_alias=True))
        print('steer.final.status:', steer_completed_status)
        print('steer.events.count:', steer_event_count)
        print('steer.assistant.preview:', steer_preview)
        print('interrupt.result:', interrupt_result.model_dump(mode='json', by_alias=True))
        print('interrupt.final.status:', interrupt_completed_status)
        print('interrupt.events.count:', interrupt_event_count)
        print('interrupt.assistant.preview:', interrupt_preview)


await async_stream_demo()

```

--------------------------------

### Parse and render templates using Template and render

Source: https://github.com/openai/codex/blob/main/codex-rs/utils/template/README.md

Demonstrates parsing a template with placeholder substitutions and literal braces, rendering it with key-value pairs, and using the one-shot render function.

```rust
use codex_utils_template::Template;
use codex_utils_template::render;

let template = Template::parse(
    "Hello, {{ name }}.\nLiteral braces: {{{{ and }}}}.\nMode: {{ mode }}",
)?;

let rendered = template.render([
    ("name", "Codex"),
    ("mode", "strict"),
])?;

assert_eq!(
    rendered,
    "Hello, Codex.\nLiteral braces: {{ and }}.\nMode: strict"
);

let one_shot = render("Hi {{ who }}!", [("who", "there")])?;
assert_eq!(one_shot, "Hi there!");
# Ok::<(), Box<dyn std::error::Error>>(())
```

--------------------------------

### Initialize exec server session

Source: https://github.com/openai/codex/blob/main/codex-rs/exec-server/README.md

Initializes a client session with the exec server. The server responds with session and environment information before the client confirms initialization.

```json
{"id":1,"method":"initialize","params":{"clientName":"example-client"}}
{"id":1,"result":{"sessionId":"00000000-0000-4000-8000-000000000001","environmentInfo":{"shell":{"name":"bash","path":"/bin/bash"},"cwd":"file:///tmp"}}}
{"method":"initialized","params":{}}
```

Sunbelt Computer Software

PL/B Language Development and Support

Codex CLI