feat: Add OrcaRouter as a supported LLM provider

[clementguarino06510-glitch]

Add OrcaRouter as a supported LLM provider

The GenAI Stack connects to OpenAI through LangChain's ChatOpenAI in chains.py, and to Bedrock through ChatBedrock and Ollama through ChatOllama. OrcaRouter speaks the same OpenAI-compatible protocol on https://api.orcarouter.ai/v1, so this PR wires it in as a first-class provider using the exact same ChatOpenAI integration — no custom base URL juggling, no generic any-endpoint passthrough.

What's added

• chains.py — a new named branch in load_llm(): any LLM=orcarouter/<model> (e.g. orcarouter/fusion-mini) is routed to ChatOpenAI pointed at the OrcaRouter base URL, mirroring the existing OpenAI case directly above it.
• api.py, bot.py, pdf_bot.py — read ORCAROUTER_BASE_URL and ORCAROUTER_API_KEY from the environment and pass them into load_llm()'s config, exactly like OLLAMA_BASE_URL is threaded through today.
• docker-compose.yml — forward ORCAROUTER_BASE_URL (default https://api.orcarouter.ai/v1) and ORCAROUTER_API_KEY to the bot, pdf_bot, and api services, alongside the existing OPENAI_API_KEY / OLLAMA_BASE_URL entries.
• env.example — a new OrcaRouter section documenting both variables.
• readme.md — documented ORCAROUTER_BASE_URL / ORCAROUTER_API_KEY in the env table and added a note under LLM Configuration.

Why OrcaRouter

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.

Verification

• python -m py_compile passes on all changed Python files; docker-compose.yml parses cleanly.
• Live-tested the new load_llm() path with a real key against https://api.orcarouter.ai/v1 — LLM=orcarouter/fusion-mini returns HTTP 200 and a valid completion.

To try it: set LLM=orcarouter/fusion-mini and ORCAROUTER_API_KEY=sk-orca-... in your .env, then docker compose up.

---

I'm an engineer on the OrcaRouter team.

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