Status: [UNTESTED] Endpoints found: http://localhost:1416/v1/chat/completions http://localhost:1416/v1/models Source: https://github.com/kaivyy/lettaproxy # Letta Proxy OpenAI-compatible proxy for [Letta Cloud](https://www.letta.com/) agents — exposes your Letta agents as standard OpenAI API endpoints with full tool-calling support and a built-in web dashboard. ## Features - **OpenAI-compatible API** — `/v1/models` and `/v1/chat/completions` endpoints work with any OpenAI SDK client - **Full tool-calling support** — bidirectional tool call round-trips (agent → caller → agent) - **Web dashboard** — manage API keys, view agents, and test chat directly from the browser - **Two proxy modes:** - `letta_proxy_dashboard.py` — full-featured with web UI - `letta_proxy_light.py` — lightweight API-only (no dashboard) - **Drop-in replacement** — point any OpenAI-compatible app (OpenClaw, LangChain, etc.) at this proxy ## Architecture ``` ┌──────────┐ OpenAI API ┌──────────────────┐ Letta SDK ┌─────────────┐ │ Client │ ── /v1/chat/* ──→│ Letta Proxy │ ── messages.* ──→│ Letta Cloud │ │ (OpenClaw│ ←── OpenAI JSON ─│ (port 1416) │ ←── stream ──── │ (agents) │ │ etc.) │ │ │ │ │ └──────────┘ └──────────────────┘ └─────────────┘ ``` ## Quick Start ### 1. Install dependencies ```bash # Using uv (recommended) uv sync # Or pip pip install fastapi uvicorn letta-client ``` ### 2. Set your Letta API key ```bash # Option A: Environment variable export LETTA_API_KEY="sk-letta-..." # Option B: File (proxy reads from .letta_key in project root) echo "sk-letta-..." > .letta_key # Option C: Via the web dashboard (POST /api/set-key) ``` ### 3. Run ```bash # Dashboard mode (recommended) python letta_proxy_dashboard.py # Li