{"id":"1474165d-f83e-4ef9-833b-b15a35dda312","slug":"enconvert","kind":"mcp","name":"enconvert","summary":"Open-source API to convert URLs and documents into PDFs, screenshots, Markdown, and structured data.","owner":"enconvert","repoUrl":"https://github.com/enconvert/enconvert","license":"AGPL-3.0","licenseTier":"strong-copyleft","trust":"community","stars":7,"match":null,"fewStars":false,"format":"mcp","createdAt":"2026-07-20T15:41:52.000Z","addedAt":"2026-10-21T04:50:53.512Z","logoUrl":"https://avatars.githubusercontent.com/u/249222219?v=4","linkOnly":false,"collection":null,"skillPath":null,"categories":["files-documents","browser-web","search"],"tags":["ai-agents","llm-tools","markdown","pdf-conversion","rag","web-scraping"],"runtime":"binary","local":true,"cloudCapable":false,"cloudCapability":"unknown","cloudReadyOnPlatform":false,"needsApiKey":false,"reviewed":false,"syncedAt":"2026-10-21T04:51:14.885Z","sample":false,"url":"https://conduits.dev/tools/enconvert","language":"Python","transports":["stdio"],"install":{},"requirements":{"env":[],"upstreamMayCharge":false},"tools":null,"readme":"

EnConvert

\n\n

\n HTTP API for converting files and reading web pages.
\n Accepts a URL or a file and returns Markdown, JSON, HTML, a screenshot, or a PDF. Supports 46 file formats.\n

\n\n

\n \"License:\n \"Release\"\n \"GitHub\n

\n\n---\n\nEnConvert reads two kinds of input through one API: files, which it converts between formats, and web pages, which it renders and returns as text or images.\n\nThis repository is the API under AGPL-3.0: request routing, the conversion pipelines, storage, background jobs, and a self-hostable render path. The hosted service at [enconvert.com](https://enconvert.com) runs this code together with additional closed components — the multi-engine render ladder, anti-bot rendering, render-quality scoring, LLM-based extraction, and semantic diffing. The table under [Open source vs Cloud](#open-source-vs-cloud) lists exactly which parts are in this repository and which are not.\n\n## Endpoints\n\n| Verb | | What it does |\n|---|---|---|\n| **perceive** | `URL → MD/JSON/PNG/PDF` | Renders a URL once and returns Markdown, HTML, a screenshot, a PDF, extracted links, and structured data. |\n| **convert** | `FILE → FILE` | Converts between 46 file types, including PDF, DOCX, XLSX, PPTX, HEIC, SVG, and CSV. |\n| **discover** | `SITE → URL MAP` | Lists a site's URLs from its sitemap and an HTTP crawl, without rendering pages. |\n| **lookup** | `QUERY → RESULTS` | Queries web, news, scholar, and maps search. |\n| **distill** | `SCHEMA → JSON` | Extracts structured data from a page against a schema you supply. |\n| **ingest** | `SITE → CHUNKS` | Crawls a site and returns JSONL chunks suitable for RAG indexing. |\n| **watch** | `PAGE → DIFF` | Checks a page on a schedule and sends a webhook when it changes. |\n\n## Quick start (hosted)\n\nCreate an API key at [enconvert.com](https://enconvert.com), then:\n\n```bash\n# perceive: a live page → Markdown\ncurl -X POST https://api.enconvert.com/v2/perceive \\\n -H \"X-API-Key: YOUR_API_KEY\" -H \"Content-Type: application/json\" \\\n -d '{\"url\": \"https://example.com\", \"outputs\": [\"markdown\"]}'\n```\n\n```bash\n# convert: a URL → PDF\ncurl -X POST https://api.enconvert.com/v1/convert/url-to-pdf \\\n -H \"X-API-Key: YOUR_API_KEY\" -H \"Content-Type: application/json\" \\\n -d '{\"url\": \"https://example.com\"}'\n```\n\n## Self-hosting\n\n```bash\ngit clone https://github.com/enconvert/enconvert.git\ncd enconvert\ndocker compose up\n```\n\nThe self-hosted build performs the file conversions and renders pages with a plain headless browser. See **[SELF_HOST.md](./SELF_HOST.md)** for the full capability list and configuration.\n\n## Render quality scoring\n\nA page fetch can succeed at the HTTP level and still return something other than the page — an anti-bot challenge, a login wall, a cookie interstitial, or an unhydrated SPA shell. Downstream consumers usually cannot tell the difference from the response body alone.\n\nThe hosted service scores every read as `render_quality` (0.0–1.0) and flags a blocked or login-walled result rather than returning it as ordinary content. This scoring is not part of the open build; a self-hosted render returns whatever the browser produced, unscored.\n\n## Open source vs Cloud\n\n| | Self-hosted (this repo) | Cloud ([enconvert.com](https://enconvert.com)) |\n|---|:---:|:---:|\n| 46-format file conversion | ✅ | ✅ |\n| perceive: URL → Markdown / screenshot / PDF | basic render | **Chromium engine, scored** |\n| discover: site → URL map | ✅ | ✅ |\n| lookup: live web search | results only | **+ render, extract & cited answers** |\n| ingest: site → RAG chunks | ✅ | ✅ |\n| watch: change monitoring | content-hash | **semantic diff** |\n| Anti-bot / stealth rendering | — | ✅ |\n| `render_quality` read scoring | — | ✅ |\n| distill: schema → structured JSON | CSS pass | **+ LLM extraction** |\n| Managed dashboard, teams, billing | — | ✅ |\n\nRows marked — are absent from this repository. The self-hosted build is functional without them: it converts files, maps sites, crawls to chunks, and renders pages with a plain headless browser.\n\n## SDKs\n\nOfficial SDKs are MIT-licensed and default to the cloud API:\n\n- [Node / TypeScript SDK](https://github.com/enconvert/node-sdk)\n- [Model Context Protocol server](https://github.com/enconvert/mcp) — one-command install for Claude, Cursor, and Windsurf\n\n## License\n\nEnConvert is open source under **AGPL-3.0** — see [LICENSE](./LICENSE). The SDKs and MCP server are **MIT**. The hosted service at [enconvert.com](https://enconvert.com) includes the additional components listed above.\n\n---\n\n

\n Pst — if EnConvert is useful to you, consider leaving a ⭐. It genuinely helps.\n

\n"}