Find the LLM calls in your repo.
Paste a public GitHub repository. Offramp downloads its Python files, reads them without running them, and lists every OpenAI, Anthropic, LiteLLM and LangChain call it finds, with the call's shape and how its list price compares with Nemotron's. No key, no sign-in, no model calls.
A full run measures parity; this page only finds call sites.
A full run generates test cases for each call, scores Nemotron against reference answers and opens a pull request for the calls that pass. It needs Docker and your own Token Factory key, so this site does not host it.
Watch the recorded demo runOr scan a local checkout:
uvx --from git+https://github.com/vj-io/offramp offramp scan .