{"repo":"sure-scale/doc-haus","free":true,"listed":false,"github":"https://github.com/sure-scale/doc-haus","clone":"git clone https://github.com/sure-scale/doc-haus.git","description":"The open-source legal AI agent. Documents stay on your machine; redlines land in Word. Fork of OpenCode.","language":"TypeScript","stars":68,"topics":["ai-agents","docx","legal-documents","legal-tech","llm","opencode","redlining","self-hosted"],"license":null,"category":"ai-agents","readme_excerpt":"doc.haus Not a chatbot. A legal team that works on your machine. Open-source multi-agent legal AI — your documents stay on your machine; the redlines land in Word. doc.haus — the website, with video demos of everything below. --- - Documents never leave your infrastructure. Each matter gets its own local index on your own disk; there is no doc.haus cloud, and the LLM models the agents run on can be local too. - Word-native. Real tracked changes baked into the .docx itself — open the result in Word and accept or reject each change, exactly as if a colleague had marked it up. - Every answer cites the clause it came from , quoted verbatim, so you can verify it before you rely on it. Get started You need Bun ( curl -fsSL https://bun.sh/install bash ). Then: The browser opens with a demo matter ready: a fictional letter of engagement, already ingested. On first run the app asks you to connect a model provider in Settings — an API key (Anthropic, OpenAI, …), a Google Cloud sign-in, or a fully local endpoint (Ollama, vLLM, LM Studio). See docs/providers.md for copy-paste configs. Try asking: What is the cap on the firm's liability? — you get the answer cited to the exact clause, quoted from the document. More things to try in demo/README.md. To work on your own documents: create a matter, upload a .docx or .pdf (including flat scans — OCR runs automatically when Poppler and Tesseract are installed), and ask. Everything — conversations, documents, the index — persists on your machine","default_branch":null,"files":null,"tree":[],"storefront":"/r/sure-scale","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/sure-scale/doc-haus/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}