Product API — Introduction¶
The Product Backend exposes a set of REST endpoints for AI evaluation, compliance, audit, and regulatory management. It runs separately from the gateway (typically on port 8199) and does not need to be in the real-time request path.
The product backend is designed to be deployed without a GPU for compliance-only use cases. When MODEL_HOST_PATH is not set, the evaluate endpoint is unavailable, but all compliance, audit, FRIA, oversight, dashboard, and reports endpoints work normally.
Base URL¶
In production behind nginx, all /v1/* paths are typically proxied to the product backend.
Route Groups¶
| Prefix | Description |
|---|---|
/glad/evaluate | Direct AI evaluation (generate + score in one call) |
/glad/export_audit | Export audit bundle for regulatory submission |
/glad/finetune | Fine-tuning job management |
/v1/ | OpenAI-compatible chat (with full evaluate pipeline) |
/v1/glad/ | Compliance, audit, FRIA, oversight, kill switch, models, etc. |
/v1/agent-flow/ | Multi-agent pipeline debugger |
/health | Health check |
OpenAI-Compatible Endpoint¶
The product backend also exposes an OpenAI-compatible /v1/chat/completions endpoint. Unlike the gateway, this endpoint runs the full evaluate pipeline inline — it generates a response with the loaded model and scores it in the same HTTP call. It accepts all the same extensions as the gateway endpoint, plus additional explain and credit-tier parameters.
See Evaluate Endpoint for a full reference.