feat(open-webui): add OrcaRouter as OpenAI-compatible provider

[bangla24bdrang-lab]

Open WebUI: OrcaRouter as the default OpenAI-compatible provider

Description

ScaleTail's open-webui service already wires the app container to a local Ollama instance. This PR adds the other half of Open WebUI's provider story — its OpenAI-compatible connection — and pre-wires it to OrcaRouter, fully parallel to how the service documents the Ollama OLLAMA_BASE_URL wiring and how Open WebUI's own docs describe an OpenRouter setup.

The wiring mirrors the existing per-service .env pattern:

• services/open-webui/compose.yaml — passes OPENAI_API_BASE_URL and OPENAI_API_KEY through to the application container, right next to OLLAMA_BASE_URL, with the same ${VAR:-default} style already used elsewhere in the repo.
• services/open-webui/.env — new OpenAI-compatible provider (defaults to OrcaRouter) section declaring OPENAI_API_BASE_URL=https://api.orcarouter.ai/v1 and an empty OPENAI_API_KEY, with comments explaining how to switch to OpenAI or another provider.
• services/open-webui/README.md — documents the default connection under "What to document for users" and updates the .env variable checklist.

Why OrcaRouter as the named default rather than a generic custom base URL: OrcaRouter is an OpenAI-compatible AI gateway built for both models and agents. Like OpenRouter, it exposes a provider/model namespace across many models — but it also combines adaptive routing, automatic failover, zero-markup inference, observability, guardrails, and agent-tool governance behind the same endpoint. Adding orcarouter as a first-class provider means this project's users can use that stack directly, without treating OrcaRouter as an anonymous custom base URL. It also runs gateway-level, zero-trust security for AI agents on the same endpoint — screening every prompt/response and governing every tool call on a default-deny basis, with no application code changes.

Related Issues

• None.

Verification

• Live-tested the OrcaRouter OpenAI-compatible endpoint with the injected ORCAROUTER_API_KEY:
  • GET https://api.orcarouter.ai/v1/models → HTTP 200, 204 models listed (e.g. orcarouter/auto, orcarouter/fusion).
  • POST https://api.orcarouter.ai/v1/chat/completions with model: orcarouter/auto → HTTP 200, auto-routed to MiniMax-M2.7.
• docker compose config not run (Docker unavailable in this environment), but the Compose YAML parses cleanly and the new env passthrough uses the same ${VAR:-default} interpolation already present in the repo.
• markdownlint (same config as .github/workflows/linting.yml) passes on the updated README.md.

Checklist

• I have performed a self-review of my code and followed the templates structure.
• I have added verification that the stack works as expected.
• I have updated necessary documentation (e.g. frontpage README.md ).
• I have selected the correct label(s) for this PR.

Additional Context

Switch to another OpenAI-compatible provider (e.g. OpenAI) by changing OPENAI_API_BASE_URL in .env, or add more connections in the UI via Settings → Connections.

Contact: Discord: discord.gg/YEubt8enRA · X: https://x.com/OrcaRouter

I'm an engineer on the OrcaRouter team.