Supported providers
FlowLens talks to models through any OpenAI-compatible or Anthropic-compatible endpoint, including providers you host yourself. The provider choice is yours, per team.
Cloud providers
OpenAI and Anthropic work out of the box: paste an API key, pick a model, and FlowLens uses it for summaries, diffs, reviews, and proposed fixes.
OpenAI-compatible endpoints
Any OpenAI-compatible base URL works: hosted gateways, proxies, or local runtimes. Set the endpoint, key, and model name, and FlowLens sends requests there instead of a first-party API.
Self-hosted models
Pair FlowLens with a model running on your own hardware and nothing leaves your network. See Self-hosting & BYOK for the full path.
What we never require
Nothing requires a shared, FlowLens-operated AI service. Your workflow data goes to the provider you configured, and your provider bills you directly.