feat(copilot): add GitHub Copilot CLI usage adapter#1023
Conversation
Read Copilot OpenTelemetry JSONL exports from the default ~/.copilot/otel directory and from COPILOT_OTEL_FILE_EXPORTER_PATH. The adapter prefers chat spans, falls back to inference and agent-turn records, and suppresses lower-priority duplicates by trace or response id. It normalizes cached input tokens out of input totals, keeps reasoning tokens in total usage, and calculates costs from model pricing rather than trusting exported Copilot cost fields. Wire the copilot namespace into focused daily, monthly, and session commands, all-source detection, and source-scoped config merging.
Regenerate the published ccusage configuration schema so the new copilot namespace is accepted by editors and config validation tooling.
Add a focused Copilot data source guide covering local OpenTelemetry JSONL export setup, report commands, calculated fields, environment variables, and troubleshooting. Update the docs navigation, source tables, configuration examples, and landing copy so the new copilot namespace is discoverable alongside the other local data sources.
Deploying with
|
| Status | Name | Latest Commit | Preview URL | Updated (UTC) |
|---|---|---|---|---|
| ✅ Deployment successful! View logs |
ccusage-guide | a523c90 | Commit Preview URL Branch Preview URL |
May 19 2026, 11:33 AM |
ccusage
@ccusage/ccusage-darwin-arm64
@ccusage/ccusage-darwin-x64
@ccusage/ccusage-linux-arm64
@ccusage/ccusage-linux-x64
@ccusage/ccusage-win32-arm64
@ccusage/ccusage-win32-x64
commit: |
ccusage performance comparisonThis compares the PR build against the base branch build on the same CI runner. Committed fixture performanceCommitted small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage. Fixtures: Claude
Large real-world-shaped fixture performanceGenerated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures. Fixtures: Claude
Package size
Lower medians and smaller packed package sizes are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
apps/ccusage/config-schema.json (1)
635-781: ⚡ Quick winVerify copilot namespace structure against established agent patterns.
All existing agents (claude, codex, opencode, amp, pi) follow a consistent structure with a top-level namespace containing
defaultsandcommandssub-properties:{ "claude": { "defaults": { ... }, "commands": { "daily": {...}, "monthly": {...} } } }The copilot commands are structured differently - they're added directly to the root-level
commandsobject with prefixed names (copilot:daily,copilot:monthly,copilot:session) and there's no separatecopilotnamespace withdefaults.This structural difference means users cannot configure copilot defaults like
copilot.defaults.offline, and must instead configure each command individually. Verify whether:
- A
copilotnamespace should exist (similar to lines 787-2314 for other agents)- The prefixed-command approach is intentional
Run the following script to check if the copilot namespace pattern is defined elsewhere or if this divergent structure is intentional:
#!/bin/bash # Check if copilot namespace structure exists elsewhere in the codebase echo "=== Searching for copilot namespace definitions in config/schema files ===" rg -n -C3 '"copilot"\s*:\s*\{' --type=json --type=ts --type=js echo "" echo "=== Searching for copilot.defaults or copilot.commands references ===" rg -n 'copilot\.defaults|copilot\.commands' --type=ts --type=js echo "" echo "=== Checking config loader for copilot namespace handling ===" rg -n -C5 'copilot.*namespace|namespace.*copilot' apps/ccusage/src/config-loader-tokens.ts echo "" echo "=== Comparing with other agent namespace patterns ===" rg -n '"(claude|codex|opencode)"\s*:\s*\{' apps/ccusage/config-schema.json | head -3🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/ccusage/config-schema.json` around lines 635 - 781, The copilot command schemas are defined as root keys ("copilot:daily", "copilot:monthly", "copilot:session") which breaks the established agent pattern of a top-level "copilot" namespace with "defaults" and "commands"; change the schema so there is a single "copilot" object containing a "defaults" property (including shared flags like offline, color/noColor, compact, json, timezone, since/until defaults) and a "commands" object that nests "daily", "monthly", and "session" schemas (move the existing "copilot:daily"/"copilot:monthly"/"copilot:session" definitions into copilot.commands.daily|monthly|session and remove the prefixed root keys), keep "additionalProperties": false and mirror other agents' structure to restore support for copilot.defaults.* configuration.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@docs/guide/environment-variables.md`:
- Around line 7-17: The env-debug grep example omitted the Copilot variable, so
update the debug snippet that filters environment variables to include COPILOT
(specifically COPILOT_OTEL_FILE_EXPORTER_PATH) so users see the Copilot export
path; locate the env-debug/diagnostic shell snippet in the same document (the
grep command around the "env-debug" example and the repeated snippet at lines
referenced 25-29) and add COPILOT to the grep pattern (e.g., include "COPILOT"
alongside existing tokens) so the COPILOT_OTEL_FILE_EXPORTER_PATH entry appears
in output when present.
---
Nitpick comments:
In `@apps/ccusage/config-schema.json`:
- Around line 635-781: The copilot command schemas are defined as root keys
("copilot:daily", "copilot:monthly", "copilot:session") which breaks the
established agent pattern of a top-level "copilot" namespace with "defaults" and
"commands"; change the schema so there is a single "copilot" object containing a
"defaults" property (including shared flags like offline, color/noColor,
compact, json, timezone, since/until defaults) and a "commands" object that
nests "daily", "monthly", and "session" schemas (move the existing
"copilot:daily"/"copilot:monthly"/"copilot:session" definitions into
copilot.commands.daily|monthly|session and remove the prefixed root keys), keep
"additionalProperties": false and mirror other agents' structure to restore
support for copilot.defaults.* configuration.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: de4ab010-9112-46e4-8755-971df8a2a332
📒 Files selected for processing (20)
apps/ccusage/config-schema.jsonapps/ccusage/src/adapter/copilot/index.tsapps/ccusage/src/adapter/copilot/parser.tsapps/ccusage/src/adapter/copilot/paths.tsapps/ccusage/src/adapter/copilot/pricing.tsapps/ccusage/src/adapter/index.tsapps/ccusage/src/adapter/types.tsapps/ccusage/src/commands/index.tsapps/ccusage/src/config-loader-tokens.tsdocs/.vitepress/config.tsdocs/guide/all-reports.mddocs/guide/claude/index.mddocs/guide/codex/index.mddocs/guide/config-files.mddocs/guide/configuration.mddocs/guide/copilot/index.mddocs/guide/environment-variables.mddocs/guide/getting-started.mddocs/guide/index.mddocs/index.md
Mention GitHub Copilot CLI in the package README source summary, source-focused examples, and unified report feature list.
Add the Copilot OpenTelemetry variable family to the environment-variable grep example so the debugging snippet matches the documented Copilot source support.
|
@cubic-dev-ai review it |
@ryoppippi I have started the AI code review. It will take a few minutes to complete. |
|
@coderabbitai review it |
|
✅ Actions performedReview triggered.
|
Replace the inline agent label conditional in the schema generator with a small exhaustive switch helper. This keeps future agent-specific labels out of nested ternaries and makes the config schema label rules easier to extend.
|
@coderabbitai review it |
|
@cubic-dev-ai review it |
@ryoppippi I have started the AI code review. It will take a few minutes to complete. |
|
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@apps/ccusage/scripts/generate-json-schema.ts`:
- Around line 160-170: Add the missing 'copilot' agent to the AGENT_NAMES
constant and update the getAgentConfigLabel function to return the correct label
for copilot: add 'copilot' to the list of AGENT_NAMES (where other agent names
like 'amp', 'claude' are defined) and add a case 'copilot': return 'copilot';
branch inside getAgentConfigLabel so the schema generator recognizes copilot
prefixes and creates the copilot namespace.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 75e92c65-97e3-4d60-81a6-353d1e82daee
📒 Files selected for processing (1)
apps/ccusage/scripts/generate-json-schema.ts
Register Copilot in the schema generator agent list and generated config schema so copilot command defaults live under the copilot namespace instead of top-level copilot:* command keys.
|
@coderabbitai review it |
|
@cubic-dev-ai review it |
@ryoppippi I have started the AI code review. It will take a few minutes to complete. |
|
✅ Actions performedReview triggered.
|
|
@coderabbitai review Updated this PR to merge current main and port GitHub Copilot CLI usage support to the Rust CLI implementation. Local validation passed:
|
|
✅ Actions performedReview triggered.
|
There was a problem hiding this comment.
Actionable comments posted: 2
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@rust/crates/ccusage/src/adapter/all.rs`:
- Around line 332-333: The table's total column is undercounting because it
recalculates totals from individual fields instead of using
summary.total_tokens(); update the table-building logic (where
input/output/cache token sums are combined) to use the already-computed
total_tokens (from summary.total_tokens()) for the displayed total, and remove
or reconcile the old recomputation (e.g., any vars like
input_tokens/output_tokens/cache_tokens used only to compute the table total) so
the table total matches the JSON summary.total_tokens().
In `@rust/crates/ccusage/src/output.rs`:
- Around line 83-84: The table footer currently sums input + output +
cache_create + cache_read but omits the new extra tokens field, causing a
mismatch with the JSON "totalTokens" (which uses input + output + cache_create +
cache_read + extra); update the non-compact footer calculation to include each
row's extra_total_tokens (or the local variable `extra`) when aggregating totals
(e.g., in the code that iterates over `rows` / `rows.iter()` and computes
`total_tokens`/footer totals) so the footer total matches the JSON
`totalTokens`.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 7a979980-cb50-4d11-9d79-f3801fcef001
📒 Files selected for processing (27)
apps/ccusage/CLAUDE.mdapps/ccusage/README.mdapps/ccusage/config-schema.jsonapps/ccusage/package.jsonapps/ccusage/src/cli.tsapps/ccusage/src/index.tsapps/ccusage/tsdown.config.tsdocs/.vitepress/config.tsdocs/guide/environment-variables.mddocs/guide/getting-started.mdpackage.jsonrust/crates/ccusage/src/adapter/all.rsrust/crates/ccusage/src/adapter/amp.rsrust/crates/ccusage/src/adapter/copilot.rsrust/crates/ccusage/src/adapter/mod.rsrust/crates/ccusage/src/adapter/opencode/mod.rsrust/crates/ccusage/src/adapter/opencode/parser.rsrust/crates/ccusage/src/adapter/pi.rsrust/crates/ccusage/src/claude_loader.rsrust/crates/ccusage/src/cli.rsrust/crates/ccusage/src/config.rsrust/crates/ccusage/src/config_schema.rsrust/crates/ccusage/src/main.rsrust/crates/ccusage/src/output.rsrust/crates/ccusage/src/progress.rsrust/crates/ccusage/src/summary.rsrust/crates/ccusage/src/types.rs
💤 Files with no reviewable changes (3)
- apps/ccusage/src/index.ts
- apps/ccusage/CLAUDE.md
- apps/ccusage/tsdown.config.ts
✅ Files skipped from review due to trivial changes (6)
- rust/crates/ccusage/src/adapter/mod.rs
- rust/crates/ccusage/src/claude_loader.rs
- apps/ccusage/README.md
- docs/.vitepress/config.ts
- docs/guide/environment-variables.md
- docs/guide/getting-started.md
| let total_tokens = summary.total_tokens(); | ||
| if total_tokens == 0 { |
There was a problem hiding this comment.
Align table total-token math with summary.total_tokens()
Line 332 now uses summary.total_tokens(), but the table still recomputes totals from only input/output/cache fields. This causes table totals to undercount whenever extra tokens exist, diverging from JSON totals.
🔧 Proposed fix
fn table_total_tokens(row: &AllRow) -> u64 {
- row.input_tokens
- .saturating_add(row.output_tokens)
- .saturating_add(row.cache_creation_tokens)
- .saturating_add(row.cache_read_tokens)
+ row.total_tokens
}📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| let total_tokens = summary.total_tokens(); | |
| if total_tokens == 0 { | |
| fn table_total_tokens(row: &AllRow) -> u64 { | |
| row.total_tokens | |
| } |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@rust/crates/ccusage/src/adapter/all.rs` around lines 332 - 333, The table's
total column is undercounting because it recalculates totals from individual
fields instead of using summary.total_tokens(); update the table-building logic
(where input/output/cache token sums are combined) to use the already-computed
total_tokens (from summary.total_tokens()) for the displayed total, and remove
or reconcile the old recomputation (e.g., any vars like
input_tokens/output_tokens/cache_tokens used only to compute the table total) so
the table total matches the JSON summary.total_tokens().
| "totalTokens": input + output + cache_create + cache_read + extra, | ||
| "totalCost": rows.iter().map(|row| row.total_cost).sum::<f64>(), |
There was a problem hiding this comment.
Fix footer total-tokens mismatch with new extra_total_tokens semantics.
After this change, JSON totals include extra tokens, but the non-compact table footer still excludes them, so totals diverge from row totals for Copilot data.
💡 Proposed fix
@@
- let total_cost = totals
+ let total_cost = totals
.get("totalCost")
.and_then(Value::as_f64)
.unwrap_or_default();
+ let total_tokens = totals
+ .get("totalTokens")
+ .and_then(Value::as_u64)
+ .unwrap_or_default();
@@
- color(
- shared,
- format_number(input + output + cache_create + cache_read),
- Color::Yellow,
- ),
+ color(shared, format_number(total_tokens), Color::Yellow),🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@rust/crates/ccusage/src/output.rs` around lines 83 - 84, The table footer
currently sums input + output + cache_create + cache_read but omits the new
extra tokens field, causing a mismatch with the JSON "totalTokens" (which uses
input + output + cache_create + cache_read + extra); update the non-compact
footer calculation to include each row's extra_total_tokens (or the local
variable `extra`) when aggregating totals (e.g., in the code that iterates over
`rows` / `rows.iter()` and computes `total_tokens`/footer totals) so the footer
total matches the JSON `totalTokens`.
ccusage performance comparisonPR SHA: This compares the Rust PR release binary against the configured base package on the same CI runner. Package runner startupExecution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one
Cached bunx execution performanceRuns the same large fixture through Fixtures: Claude
Package runtime diagnosticsCompares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself. Fixtures: Claude
Committed fixture performanceCommitted small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage. Fixtures: Claude
Large real-world-shaped fixture performanceGenerated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures. Fixtures: Claude
Artifact size
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees. |
ccusage performance comparisonPR SHA: This compares the PR package against the configured base package on the same CI runner. Package runner startupExecution setup measures any pre-benchmark package materialization used by the execution benchmark. Bunx temp cache measures one
Cached bunx execution performanceRuns the same large fixture through Fixtures: Claude
Package runtime diagnosticsCompares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself. Fixtures: Claude
Committed fixture performanceCommitted small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage. Fixtures: Claude
Large real-world-shaped fixture performanceGenerated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures. Fixtures: Claude
Artifact sizeLower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees. |
1 participant
Try This PR
Summary
Verification
Summary by cubic
Adds a GitHub Copilot CLI usage adapter, now implemented in Rust, that reads local OpenTelemetry JSONL exports, integrates with unified reports, and provides daily/monthly/session views with pricing and token accounting that include reasoning tokens.
New Features
copilotadapter reads~/.copilot/otel/*.jsonlandCOPILOT_OTEL_FILE_EXPORTER_PATH.ccusage copilot daily|monthly|session. Included in all-source detection, progress labels, andcopilot-scoped config.Docs
copilotnamespace with per-command overrides.Written for commit a523c90. Summary will update on new commits. Review in cubic
Summary by CodeRabbit
New Features
Documentation