{
  "markdown": "[![Frenchie — your agent's best friend.](https://getfrenchie.dev/brand/frenchie-readme-wordmark.svg)](https://getfrenchie.dev)\n\n[![smithery badge](https://smithery.ai/badge/lab94/frenchie)](https://smithery.ai/servers/lab94/frenchie)\n\n# @lab94/frenchie\n\n**Frenchie — your agent's best friend.**\n\nInstall Frenchie in your coding agent with one command. Frenchie Kit gives agents file and multimodal capabilities: read PDFs and images, transcribe audio and video, extract Office/CSV files to Markdown, and generate images from text prompts. Frenchie Method gives teams a structured workflow from requirement to release.\n\n> **Upgrading from 0.1.x or 0.2.x?** See [MIGRATION.md](./MIGRATION.md) for the breaking changes in 0.3.0 (stdio metadata-only responses, absolute-path MCP configs, new `mcp --help` / `--selftest` flags).\n\nThis package ships:\n\n- Frenchie **Kit** (`/ocr`, `/transcribe`, `/extract`, `/generate-image`, `/frenchie-status` commands + HTTP/stdio MCP guidance)\n- Frenchie **Method** (`/frenchie-method`, `/requirement`, `/qa-analysis`, `/implement`, `/run-test`, `/pr-review`, `/release-check`, and related workflow commands)\n- The Frenchie **stdio MCP server** (`lab94-frenchie mcp`) bundled for `npx`\n- An **installer** that wires skills and MCP config into your agent\n\n## Quick start\n\n### 1. Install the skills\n\nFrom the root of your project:\n\n```bash\nnpx @lab94/frenchie install\n```\n\nThis installs Frenchie Kit and Frenchie Method skills. Method is text-only and works without an API key.\n\nOptional Method config:\n\n```bash\nnpx @lab94/frenchie install --init-method\n```\n\nThis creates `.frenchie/method.config.json` and `.frenchie/.gitignore`; it does not edit your root `.gitignore`.\n\n### 2. Log in for Kit MCP tools when you need file or image capabilities\n\nCreate an account at [getfrenchie.dev](https://getfrenchie.dev). You get **100 free credits on your first signup, once per email**. No card required.\n\nThen connect your local agent:\n\n```bash\nnpx @lab94/frenchie login\n```\n\nThe login command opens a browser device-code flow, then writes `FRENCHIE_API_KEY` into the MCP config the installer created. The full key is never printed.\n\nFallback if you already have a key from the dashboard:\n\n```bash\nnpx @lab94/frenchie install --api-key fr_your_key_here\n```\n\nTo target a specific agent:\n\n```bash\nnpx @lab94/frenchie login --agent claude\nnpx @lab94/frenchie login --agent cursor\nnpx @lab94/frenchie login --agent codex\nnpx @lab94/frenchie login --agent vscode\nnpx @lab94/frenchie login --agent gemini\n```\n\nUser-level installs (Antigravity, Windsurf, Zed, Claude Desktop) need the `--global` flag — auto-detect never writes under `$HOME`:\n\n```bash\nnpx @lab94/frenchie login --agent antigravity   --global\nnpx @lab94/frenchie login --agent windsurf       --global\nnpx @lab94/frenchie login --agent zed            --global\nnpx @lab94/frenchie login --agent claude-desktop --global\n```\n\nFallback with an existing key:\n\n```bash\nnpx @lab94/frenchie install --agent claude        --api-key fr_…\nnpx @lab94/frenchie install --agent codex         --api-key fr_…\nnpx @lab94/frenchie install --agent antigravity   --global --api-key fr_…\nnpx @lab94/frenchie install --agent claude-desktop --global --api-key fr_…\n```\n\nAfter login, your agent can call `ocr_to_markdown`, `transcribe_to_markdown`, `extract_to_markdown`, or `generate_image`. OCR/transcription/extraction results are saved to `.frenchie/<name>/result.md` automatically; generated images are saved to `.frenchie/<slug>/generated.<ext>`.\n\n### Upgrading later\n\nWhen a new Frenchie version ships, the pinned package spec inside each MCP config (`@lab94/frenchie@0.5.0`) needs to point at the new version so `npx` stops serving a cached older bundle. Two upgrade commands cover everything:\n\n```bash\n# Project-scoped configs in this directory (.mcp.json, .cursor/mcp.json, .codex/config.toml, etc.)\nnpx @lab94/frenchie@latest install\n\n# Every $HOME config that already has a frenchie entry — Antigravity, Claude Desktop, Windsurf, Zed,\n# and any global Claude Code / Cursor / Codex / Gemini configs you set up. cwd doesn't matter.\nnpx @lab94/frenchie@latest install --global\n```\n\n`install --global` is upgrade-mode: it only rewrites `frenchie` entries that already exist. It will not surprise-install anything new under `$HOME`. First-time global installs still go through `install --agent <name> --global`.\n\n### 3. Restart your agent\n\nThe installer prints the restart hint for your agent. After that, ask:\n\n```\nOCR ./report.pdf with Frenchie\n```\n\n…and Frenchie takes it from there.\n\nFor workflow orchestration, ask:\n\n```txt\n/frenchie-method\n```\n\n`/frenchie` remains the Kit entrypoint in agents that use a single Frenchie command. It is not an alias for Method.\n\n## Hosted agents (Lovable, Manus, Claude.ai, ChatGPT.com, Le Chat)\n\nThese agents can't run local npm binaries. Use the hosted MCP endpoint instead:\n\n```\nURL:    https://mcp.getfrenchie.dev\nHeader: Authorization: Bearer fr_your_key_here\n```\n\nThe same `@lab94/frenchie` skill files work in HTTP mode — install them once with `install --agent <name>` and the included SKILL.md will tell the agent to upload files via `upload_file` before calling OCR/transcription/extraction. Image generation does not need an upload step in HTTP mode; it returns a short-lived `imageUrl` that the agent should download for the user.\n\n## What you get\n\n### Frenchie Kit\n\n| Command | What it does |\n|---------|-------------|\n| `/ocr <file>` | Parse a PDF or image into Markdown |\n| `/transcribe <file>` | Parse audio or video into a Markdown transcript |\n| `/extract <file>` | Parse DOCX, XLSX, CSV, TSV, or PPTX into Markdown |\n| `/generate-image <prompt>` | Generate a single image from a text prompt |\n| `/frenchie-status` | Check credits and recent jobs |\n\nUnder the hood, Frenchie exposes these MCP tools:\n\n- `ocr_to_markdown`\n- `transcribe_to_markdown`\n- `extract_to_markdown`\n- `generate_image`\n- `get_job_result`\n- `upload_file` (HTTP mode only)\n- `fetch_result_file` (HTTP mode only)\n\n### Frenchie Method\n\nSlash notation is the canonical Method command name. Claude Code reads command\nfiles directly. Codex Desktop gets the same entries through wrapper skills such\nas `requirement` and `qa-analysis`; restart Codex after install before trying\n`/requirement`. Hosts without slash support can use natural language such as\n`Frenchie Method requirement ...`.\n\n| Command | What it does |\n|---------|-------------|\n| `/frenchie-method` | Route a task through Method from requirement to release |\n| `/method-check` | Validate Method config, artifacts, gates, and next action |\n| `/requirement` | Turn a request into scoped requirements |\n| `/qa-analysis` / `/test-case` | Design proof before implementation planning |\n| `/sa-analysis` / `/implementation-plan` | Analyze architecture and implementation approach |\n| `/implement` | Execute from Method artifacts using the current agent |\n| `/run-test` | Run repo-declared verification and write a test report |\n| `/pr-review` / `/review-pr` | Review the result, tests, security, and artifact drift |\n| `/release-check` | Read-only release readiness gate |\n\nMethod stores optional workflow artifacts under `.frenchie/docs/feature/`. `install --init-method` also writes a visible `skillTriggers` skeleton in `.frenchie/method.config.json`; command keys are empty slots teams can fill with installed agent skills, for example `\"requirement\": [\"brainstorming\"]`. Teams decide whether to commit or ignore Method config and artifacts.\n\n## Invocation — how to call Frenchie in each agent\n\nEvery agent handles MCP a little differently. `/ocr` is a Claude Code-only slash command; Method slash commands are host-dependent and use Codex wrapper skills outside Claude Code. Other agents use natural language, `@`-mention, or a server-name slash command. All facts below are dogfood-verified.\n\n| Agent | Invoke | Full guide |\n|-------|--------|------------|\n| Claude Code | `/ocr TOR.pdf` | [docs](https://getfrenchie.dev/docs/tools/claude-code) |\n| Cursor | `Use Frenchie to OCR TOR.pdf` | [docs](https://getfrenchie.dev/docs/tools/cursor) |\n| Codex (Desktop / CLI / IDE) | Kit: `/frenchie TOR.pdf` · `@frenchie ocr TOR.pdf`; Method: `/requirement` after restart, or `Frenchie Method requirement ...` | [docs](https://getfrenchie.dev/docs/tools/codex) |\n| Antigravity | `/frenchie TOR.pdf` (invokes by server name) | [docs](https://getfrenchie.dev/docs/tools/antigravity) |\n| VS Code Copilot | `/frenchie TOR.pdf` | [docs](https://getfrenchie.dev/docs/tools/vscode) |\n| Claude Desktop | `Use Frenchie to OCR TOR.pdf` | [docs](https://getfrenchie.dev/docs/tools/claude-desktop) |\n| Windsurf | `OCR TOR.pdf via Frenchie` | [docs](https://getfrenchie.dev/docs/tools/windsurf) |\n| Gemini CLI | `OCR TOR.pdf with Frenchie` | [docs](https://getfrenchie.dev/docs/tools/gemini-cli) |\n| Zed | `OCR TOR.pdf via Frenchie` | [docs](https://getfrenchie.dev/docs/tools/zed) |\n\nSomething not working? See the [symptom-first troubleshooting guide](https://getfrenchie.dev/docs/troubleshooting) — every error we've hit in dogfood has a canonical entry.\n\n## Pricing\n\nSimple numbers. No subscriptions.\n\nFile extraction pricing: DOCX 0.5 credits per page, XLSX 0.5 credits per sheet, CSV/TSV 0.5 credits per file, PPTX 1 credit per slide.\n\n| Action | Cost |\n|--------|------|\n| OCR | 1 credit per page |\n| Transcription | 2 credits per minute |\n| DOCX extraction | 0.5 credits per page |\n| XLSX extraction | 0.5 credits per sheet |\n| CSV/TSV extraction | 0.5 credits per file |\n| PPTX extraction | 1 credit per slide |\n| Image generation | 20 credits per image |\n\n**$1 = 100 credits.** Credits don't expire.\n\n## Privacy\n\nFiles are processed and deleted. Results expire about 30 minutes after first delivery. If you need a durable copy, save the Markdown when it comes back.\n\n## Supported formats\n\n**OCR:** PDF, PNG, JPG, JPEG, WebP\n\n**Transcription:** MP3, M4A, WAV, MP4, MOV, WebM\n\n**File extraction:** DOCX, XLSX, CSV, TSV, PPTX\n\n**Image generation:** PNG, JPEG, WebP output from text prompts\n\n## Need help?\n\n- [Docs](https://getfrenchie.dev/docs)\n- [Dashboard](https://getfrenchie.dev/dashboard)\n- [LAB94](https://lab94.io)\n- [support@getfrenchie.dev](mailto:support@getfrenchie.dev)\n",
  "bytes": 10215,
  "sha": "dfcd1611aef7f04595eb36af48f93d8f33ed92fc25f6b756d481e51c0f24c9a8",
  "repo_slug": "lab94/frenchie-skill",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_lab94_frenchie_skill_af54df9b/readme"
}