A self-improving AI agent for personal use. Runs in your terminal, connects via Telegram, and remembers across sessions. Use any LLM provider -- OpenRouter, z.ai/GLM, OpenAI, Anthropic, or your own endpoint. Switch with hermes model.
Fork of Hermes Agent by Nous Research, with ~30,000 lines of removed code (RL training, 13 gateway adapters, marketplace infrastructure, vendor-specific routing). See docs/roadmap.md for the full simplification history.
| Terminal interface | Full TUI with multiline editing, slash-command autocomplete, conversation history, interrupt-and-redirect, and streaming tool output. |
| Telegram + API gateway | Telegram bot and HTTP API server from a single gateway process. Cross-platform conversation continuity. |
| Local skills and memory | Agent-curated memory with periodic nudges. Autonomous skill creation from experience. FTS5 session search with LLM summarization for cross-session recall. Honcho dialectic user modeling. Local-only skill loading. |
| Scheduled automations | Cron scheduler (opt-in toolset). Daily reports, nightly backups, weekly audits -- all in natural language, running unattended. |
| Delegates and parallelizes | Spawn isolated subagents for parallel workstreams. Write Python scripts that call tools via RPC, collapsing multi-step pipelines into zero-context-cost turns. |
| Runs anywhere | Terminal backends: local, Docker, SSH, and Modal. Run it on a $5 VPS or a GPU cluster. |
One-line install (Linux/macOS):
curl -fsSL https://raw.githubusercontent.com/joaomj/hermes-agent-fork/main/scripts/install.sh | bashFrom source:
git clone https://github.com/joaomj/hermes-agent-fork.git
cd hermes-agent-fork
curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv .venv --python 3.11
source .venv/bin/activate
uv pip install -e ".[all]"
hermes setupAfter installation:
source ~/.zshrc # or ~/.bashrc
hermes # start chattinghermes # Interactive CLI -- start a conversation
hermes model # Choose your LLM provider and model
hermes tools # Configure which tools are enabled
hermes config edit # Open config in your editor
hermes gateway # Start the messaging gateway (Telegram + API server)
hermes setup # Full setup wizard (API keys, model, tools)
hermes doctor # Diagnose configuration issuesHermes has two entry points: start the terminal UI with hermes, or run the gateway and talk to it from Telegram or via the API server. Slash commands are shared across both.
| Action | CLI | Telegram / API |
|---|---|---|
| Start chatting | hermes |
hermes gateway start, then message the bot |
| Fresh conversation | /new or /reset |
/new or /reset |
| Change model | /model [provider:model] |
/model [provider:model] |
| Set personality | /personality [name] |
/personality [name] |
| Retry / undo | /retry, /undo |
/retry, /undo |
| Context / usage | /compress, /usage |
/compress, /usage |
| Interrupt | Ctrl+C or new message |
/stop or new message |
Upstream documentation at hermes-agent.nousresearch.com/docs covers most features in detail (CLI usage, configuration, tools, skills, memory, MCP integration, security, context files).
git clone https://github.com/joaomj/hermes-agent-fork.git
cd hermes-agent-fork
uv venv .venv --python 3.11
source .venv/bin/activate
uv pip install -e ".[all,dev]"
python -m pytest tests/ -qMIT -- see LICENSE.
Original work by Nous Research. Fork modifications under the same license.