Appearance

Agents and Client Compatibility ​

Relay is designed for local-agent workflows where tools expect hosted API shapes.

Quick Connect ​

After setup, point any OpenAI/Anthropic-compatible agent at:

Base URL: http://127.0.0.1:1234/v1
API Key:  (leave blank unless API_KEY is set in .env)
Model:    any model from /v1/models

Tested Clients ​

Relay works with any client that allows custom base URLs and model IDs:

  • opencode — set provider.base_url to http://127.0.0.1:1234/v1
  • Cursor — add as custom OpenAI-compatible provider
  • Claude Code — use ANTHROPIC_BASE_URL=http://127.0.0.1:1234/v1
  • Continue — add as OpenAI-compatible provider in config
  • Aider — --openai-api-base http://127.0.0.1:1234/v1

Session Affinity ​

For multi-project setups, use the session-id header to keep conversations isolated:

bash
# Project A
curl -H "session-id: project-alpha" http://127.0.0.1:1234/v1/chat/completions ...

# Project B — Relay restarts the model to clear context
curl -H "session-id: project-beta" http://127.0.0.1:1234/v1/chat/completions ...

Public Sharing ​

See Public Deployment to share your relay with friends over HTTPS.

Known Limits ​

  • Relay does not implement hosted platform orchestration (assistants, threads, runs)
  • Non-function tools (web search, code interpreter) are stripped
  • See API Compatibility for full endpoint support