YandexGPT Proxy
OpenAI-compatible translation proxy for Yandex Cloud Foundation Models (YandexGPT).
What it does
Runs a local HTTP proxy on port 8444 that accepts OpenAI-format API calls and translates them to the YandexGPT API. Zero external dependencies — pure Node.js.
Supported Models
| Model | YandexGPT Model URI |
|---|---|
| yandexgpt | gpt://FOLDER_ID/yandexgpt/latest |
| yandexgpt-lite | gpt://FOLDER_ID/yandexgpt-lite/latest |
| yandexgpt-32k | gpt://FOLDER_ID/yandexgpt-32k/latest |
Setup
- Get a Yandex Cloud API key and folder ID
- Save to
~/.openclaw/yandexgpt.env:YANDEX_API_KEY=your_api_key YANDEX_FOLDER_ID=your_folder_id YANDEX_PROXY_PORT=8444 - Run
scripts/setup.sh - Run
scripts/start.sh - Run
scripts/patch-config.shto add to OpenClaw config
Endpoints
GET /v1/models— List available modelsPOST /v1/chat/completions— Chat completion (OpenAI format)
Scripts
scripts/setup.sh— Create env file templatescripts/start.sh— Start proxyscripts/stop.sh— Stop proxyscripts/status.sh— Check statusscripts/patch-config.sh— Add YandexGPT to openclaw.json