This repository is the Gemini CLI
extension for the Unstructured Transform
MCP server (https://mcp.transform.unstructured.io). It lets Gemini CLI
parse documents (PDF, DOCX, PPTX, XLSX, HTML, EML, images, and multiple
other formats) into markdown, element JSON, HTML, or plain text.
The Transform documentation is the source of truth for what the server does, its supported file types, parsing options, request limits, and billing. This repository only carries the Gemini extension manifest and the agent guidance the extension ships.
gemini extensions install https://github.com/Unstructured-IO/transform-gemini-extensionThen sign in once with /mcp auth transform (browser OAuth; no API key
needed). The extension registers the transform server and ships the
agent guidance in AGENTS.md.
Using another tool (Claude Code, Claude Desktop, Cursor, Codex, Google Antigravity, Cline)? See the official install guides and the Unstructured MCP integrations index.
Ask Gemini to parse something, for example: "Parse report.pdf to
markdown." The agent uploads the file, starts a job, polls until it
finishes, and writes the structured result. Large or scanned files can
take a few minutes. AGENTS.md carries the guidance agents
follow for this flow.
├── AGENTS.md # agent guidance, shipped with the extension
├── gemini-extension.json # Gemini CLI extension manifest
├── scripts/validate.sh # consistency checks (plain bash)
├── .github/workflows/
│ └── validate.yml # runs the checks in CI
├── CODEOWNERS
├── CONTRIBUTING.md
├── LICENSE
├── README.md
├── .editorconfig
├── .gitattributes
└── .gitignore
./scripts/validate.shChecks that the manifest is valid JSON and that no tracked file
references a server URL other than the one in the manifest. Runs in CI on
pull requests and pushes to main.
This repository is maintained by the Unstructured team. Bug reports and suggestions are welcome as issues. See CONTRIBUTING.md before opening a pull request.