AssemblyAI, HoneyHive and now IMG Processing all expose search plus a read-only shell over a virtual docs filesystem plus a feedback tool. That convergence is a good sign: it means a client that learns the pattern once can use any of them, and it means grep-over-docs has been recognised as more useful than search alone. Free and read-only, so there is no reason not to connect it while building.
IMG Processing's documentation endpoint, which answered our probe without credentials as IMG Processing 1.0.0. This documents their image API rather than performing image work.
- search_img_processing — the knowledge base, including code examples and API reference
- query_docs_filesystem_img_processing — a read-only shell over a virtual filesystem of the docs
- submit_feedback — report a documentation problem with its path
None for the documentation. The API itself has a free test key, per their authentication guide.
One command — npx -y mcp-remote https://docs.img-processing.com/mcp