Chunking, a configurable memory limit and node-type filtering exist for one reason: design files that would otherwise exhaust the process or the context before you get an answer. On a small file this behaves like an ordinary Figma reader; on a large one it is the difference between an answer and a crash.
A Figma API server built for big files, with memory-aware processing, chunked retrieval and pagination on every listing call.
- File data retrieved in chunks, with page size, a memory ceiling in MB and a node-type filter set per call
- Resume for an interrupted retrieval, so a large file does not have to restart from zero
- Pagination on the listing operations (list_files)
- File metadata, components and styles (get_file_info, get_components, get_styles)
- Version history and comments on a file (get_file_versions, get_file_comments)
- Progress tracking and debug logging while a long retrieval runs
A Figma access token, given either in the environment or in a JSON file you point the server at. Built from source with npm install and npm run build, or installed for Claude Desktop through Smithery.
One command plus a key — npx -y @smithery/cli install @ArchimedesCrypto/figma-mcp-chunked --client claude, then supply credentials