mdelias opens the fork up to OpenRouter's whole model shelf

The fork could already store an OpenRouter key but had no way to actually use it - now it does.

integrationinfrastructure

OpenRouter is a single gateway that fronts hundreds of AI models, both open-source and commercial, behind one account and one key. This fork knew how to save an OpenRouter key but had no wiring to send requests through it, so the option was dead on arrival. That plumbing is now in place, sitting alongside the fork's existing Claude, Gemini, and OpenAI connections.

Both streaming replies and tool-driven, agent-style steps run through the new route, so models reached this way behave like first-class citizens rather than a bolt-on. In practice, OpenRouter's full catalogue becomes selectable inside the fork with no further setup.

So what For legal teams who want to trial cheaper or specialized models without re-plumbing their stack, this turns a locked door into a dropdown.

View this fork on GitHub →

Spotted something wrong? Or know the PR text has fresher detail than the writeup above?

Commits in this thread

1 commit from mdelias/mike, oldest first. Source extracted verbatim from the harvested git log.

SHA Subject Author Date
438b6f08 feat: add OpenRouter LLM provider support railway-app[bot] 2026-06-28 ↗ GitHub
commit body
- Add openrouter.ts with streamOpenRouter and completeOpenRouterText
- Add OPENROUTER_MAIN/MID/LOW_MODELS constants to models.ts
- Update providerForModel() to detect openrouter/ prefix
- Expand ALL_MODELS set to include OpenRouter model IDs
- Add "openrouter" to Provider union type in types.ts
- Wire OpenRouter into streamChatWithTools and completeText in index.ts

Capture this thread into my fork

Download a single Markdown prompt that tells Claude how to port every commit above into your working tree — adapting paths and structure to match your repo. Run it via claude -p < capture-thread-696.md from inside the repo you want the changes in.

⬇ Download capture-thread-696.md