Gen3 Gt Api Runbooks Open Webui - GT-Edge-AI/GT-AI-OS GitHub Wiki

Start Here

Configure Open WebUI admin connection settings with GT API base URL (OPENAI_API_BASE_URL) and a scoped inference key; published aliases appear as selectable models. Recommended preset: OpenAI-compatible chat client.

Why this matters

This runbook maps Open WebUI to GT API routes operators actually publish.

Details

Compatibility: Guided OpenAI-compatible setup ยท Category: Workflow and builders

Official documentation

Configuration fields

  • OPENAI_API_BASE_URL: https://<tenant-host>/api/tenant
  • OPENAI_API_KEY: gtak_...
  • Admin โ†’ Connections โ†’ OpenAI โ†’ API Base URL: https://<tenant-host>/api/tenant

GT route mapping

GT route Verdict Client integration
GET /v1/models native Open WebUI model refresh calls /v1/models via configured base URL
POST /v1/chat/completions native Open WebUI chat โ†’ /v1/chat/completions
POST /v1/embeddings native When embed scope on key and Open WebUI embed config uses same base
POST /v1/audio/transcriptions not_supported Not default Open WebUI OpenAI connection path
POST /v1/audio/speech not_supported Not default Open WebUI OpenAI connection path
POST /v1/images/generations not_supported Not default unless image features configured with GT image scope
POST /v1/conversations/files not_supported External script โ€” not default UI attachment pipeline
POST /v1/datasets/{id}/files not_supported External ingestion automation to GT dataset route
GET /v1/files/{id} not_supported External polling after GT upload

Not supported in this product

  • Open WebUI default file attachment flow is product-local, not GT conversation or dataset upload routes.

Prerequisites

  • Publish aliases chat users should see.
  • Create inference key with chat (and embed if needed) scopes.

Setup steps

  1. Set OPENAI_API_BASE_URL to tenant GT API URL in container env or admin UI.
  2. Paste GT API key into OpenAI connection settings.
  3. Reload models and confirm only allowlisted published aliases appear.
  4. Complete one user chat turn with a published alias.

GT extensions and caveats

  • Use separate automation for /v1/datasets/{id}/files โ€” default UI attachments may not map to GT.
  • Enable embed scope only when Open WebUI calls /v1/embeddings through GT.

Validation checklist

  • User completes chat with published alias.
  • Model picker does not leak disallowed aliases.
  • Key rotation documented for Open WebUI operators.

Plain-text export

Open WebUI runbook
Guided OpenAI-compatible setup ยท Workflow and builders
Recommended key preset: OpenAI-compatible chat client
Evidence: documented compatibility (vendor docs cross-check)

Configure Open WebUI admin connection settings with GT API base URL (`OPENAI_API_BASE_URL`) and a scoped inference key; published aliases appear as selectable models.

Official documentation:
- https://docs.openwebui.com/getting-started/env-configuration/
- https://docs.openwebui.com/getting-started/quick-start/
- https://docs.openwebui.com/features/chat-features/

Configuration fields:
- OPENAI_API_BASE_URL: https://<tenant-host>/api/tenant
- OPENAI_API_KEY: gtak_...
- Admin โ†’ Connections โ†’ OpenAI โ†’ API Base URL: https://<tenant-host>/api/tenant

GT route mapping:
- GET /v1/models (native): Open WebUI model refresh calls `/v1/models` via configured base URL
- POST /v1/chat/completions (native): Open WebUI chat โ†’ `/v1/chat/completions`
- POST /v1/embeddings (native): When embed scope on key and Open WebUI embed config uses same base
- POST /v1/audio/transcriptions (not_supported): Not default Open WebUI OpenAI connection path
- POST /v1/audio/speech (not_supported): Not default Open WebUI OpenAI connection path
- POST /v1/images/generations (not_supported): Not default unless image features configured with GT image scope
- POST /v1/conversations/files (not_supported): External script โ€” not default UI attachment pipeline
- POST /v1/datasets/{id}/files (not_supported): External ingestion automation to GT dataset route
- GET /v1/files/{id} (not_supported): External polling after GT upload

Not supported in this product:
- Open WebUI default file attachment flow is product-local, not GT conversation or dataset upload routes.

Prerequisites:
- Publish aliases chat users should see.
- Create inference key with chat (and embed if needed) scopes.

Setup steps:
1. Set `OPENAI_API_BASE_URL` to tenant GT API URL in container env or admin UI.
2. Paste GT API key into OpenAI connection settings.
3. Reload models and confirm only allowlisted published aliases appear.
4. Complete one user chat turn with a published alias.

GT extensions and caveats:
- Use separate automation for `/v1/datasets/{id}/files` โ€” default UI attachments may not map to GT.
- Enable embed scope only when Open WebUI calls `/v1/embeddings` through GT.

Validation checklist:
- User completes chat with published alias.
- Model picker does not leak disallowed aliases.
- Key rotation documented for Open WebUI operators.