EnConvert - One ingestion API for AI agents and RAG pipelines

𝐌𝐨𝐬𝐭 𝐑𝐀𝐆 𝐩𝐢𝐩𝐞𝐥𝐢𝐧𝐞𝐬 𝐛𝐫𝐞𝐚𝐤 𝐚𝐭 𝐢𝐧𝐠𝐞𝐬𝐭𝐢𝐨𝐧, 𝐧𝐨𝐭 𝐫𝐞𝐭𝐫𝐢𝐞𝐯𝐚𝐥. Three parsers, three output formats, three failure modes before a single token reaches your model. EnConvert puts one interface in front of all of it. 🔹 𝐅𝐨𝐮𝐫 𝐞𝐧𝐝𝐩𝐨𝐢𝐧𝐭𝐬 𝐥𝐢𝐯𝐞: • 𝘗𝘦𝘳𝘤𝘦𝘪𝘷𝘦: any URL to clean Markdown • 𝘐𝘯𝘨𝘦𝘴𝘵: PDF and DOCX for RAG • 𝘊𝘰𝘯𝘷𝘦𝘳𝘵: file and URL format conversion • 𝘋𝘪𝘴𝘵𝘪𝘭𝘭: summarisation in one call 🔹 MCP server: Anthropic Registry, LobeHub, Glama. 🔹 Free tier: 500 ops/month, no card. 𝐓𝐫𝐲 𝐢𝐭 𝐚𝐧𝐝 𝐭𝐞𝐥𝐥 𝐮𝐬 𝐰𝐡𝐚𝐭 𝐛𝐫𝐞𝐚𝐤𝐬.

Hey Product Hunt! 👋 Krystin here, commercial co-founder at EnConvert.

We built this because every RAG pipeline and LLM agent we looked at had the same problem underneath it: a different tool for each input type, each with its own key, its own output format, and its own way of failing. The ingestion layer was the part nobody wanted to maintain but everyone had to build.

I am the non-technical half of this team, so I will keep it plain.

What is live today: four endpoints. Perceive turns any web page into clean Markdown, including JavaScript-rendered pages. Ingest handles PDF and DOCX extraction for RAG pipelines. Convert handles file and URL format conversion. Distill summarises any content in one API call. MCP server is live on the Anthropic MCP Registry, LobeHub, and Glama. Free tier is 500 ops per month, no card required.

What is coming in Phase 2: Lookup for live web search🔍, Watch for page change monitoring👁, and Discover for web crawling🕸.

We are early. We know there are edge cases we have not hit yet. If you try it and something breaks, that is genuinely useful to us. Tell us what fails. 🙏

Krystin, Het, and Om

EnConvert - One ingestion API for AI agents and RAG pipelines

Replies