Local LLM

LLMZap Gateway

The platform now supports a configurable local OpenAI-compatible LLM gateway. Change the base URL and endpoint paths through the server environment without changing application code.

Not configured

Saved Config

Frontend-managed gateway

Saved config overrides env defaults. You can still override the connection per request for one-off testing.

Connection

Gateway configuration

Provider llmzap-openaiBase URL —Default model —Auth mode bearerSaved key —Updated 2026-03-29T11:20:39.847Z
{
  "health": "/health",
  "models": "/v1/models",
  "props": "/props",
  "chatCompletions": "/v1/chat/completions",
  "completions": "/v1/completions",
  "embeddings": "/v1/embeddings"
}

Env knobs: `LLM_API_BASE_URL`, `LLM_CHAT_COMPLETIONS_PATH`, `LLM_MODELS_PATH`, `LLM_EMBEDDINGS_PATH`, `LLM_DEFAULT_MODEL`, `LLM_API_KEY`.

Models

Remote inventory

No models reported
—

Chat Test

Direct model call

—

Planner Test

Mirror-aware plan

—