Overview
Mike — also written MikeOSS — is an open-source platform for the document-heavy parts of legal work: reviewing documents, drafting, and researching case law. It is a full application rather than a library, built from a Next.js frontend, an Express backend, Supabase for auth and Postgres, and S3-compatible object storage.
The workflow is organised the way a matter is: projects, folders and a document library, with chat over the documents in an open matter. Review runs as a workflow — reusable assistant and tabular-review workflows that you apply to a set of documents — and suggested edits come back for a person to apply. Citations are checked rather than trusted: the CourtListener integration verifies them and backs US case-law research.
Where the work actually happens is covered too. A beta Word task-pane add-in puts the assistant inside the document a lawyer is drafting in, and exports are tamper-evident. Model access is bring-your-own — an Anthropic, Gemini or OpenAI key, or supported models run locally through Ollama, which is the option that matters when the documents cannot leave the building.
The system assistant and tabular-review workflows are maintained in a separate `mike-workflows` repository and packaged into the application, so the prompt logic can be reviewed and versioned on its own.
What it does
- Chat with the documents in an open matter, over a project/folder/library structure
- Reusable assistant and tabular-review workflows for repeated document review
- Document review with suggested edits a person applies
- Citation verification and US case-law research through CourtListener
- Microsoft Word task-pane add-in (beta) for drafting in place
- Bring-your-own model keys (Anthropic, Gemini, OpenAI) or fully local inference through Ollama
- Self-hosted by design: Docker Compose stack with Supabase, RustFS object storage and local email capture
Getting started
The bundled Compose stack runs the app plus Supabase, object storage and email capture with no managed infrastructure. These are the README's local-development steps — its credentials are for local use only.
Copy the environment templates
Both the root and backend environments have examples.
cp .env.example .env
cp backend/.env.example backend/.envGenerate the two secrets
In backend/.env, set DOWNLOAD_SIGNING_SECRET and USER_API_KEYS_ENCRYPTION_SECRET to separate generated values.
openssl rand -hex 32Add a model key
Put an Anthropic, Gemini or OpenAI key in backend/.env — unless you plan to run everything through Ollama.
Start the stack
Then open the app and create an account.
docker compose up --build
# http://localhost:3000Commands and code are distilled from the project's own documentation — always check the official repo for the latest.
When to use it
- Give a small firm document review and drafting help without sending client files to a hosted vendor
- Run the same tabular review across a batch of contracts and get a comparable table back
- Check every citation in a draft against CourtListener before it goes out
- Keep an AI drafting assistant inside Word, where the document is already being written
How Mike (MikeOSS) compares
Mike (MikeOSS) alongside other open-source rag frameworks & platforms tools AI/TLDR tracks, ranked by GitHub stars.
| Tool | Stars | What it does |
|---|---|---|
| Dify | ★ 156k | An open-source platform with a visual workflow builder for creating LLM and RAG applications without writing much code. |
| graphify | ★ 119k | Turns a folder of code, docs, PDFs and images into a local knowledge graph with tree-sitter AST parsing and Leiden communities — queryable by agents over MCP, no vector store. |
| RAGFlow | ★ 90.9k | A RAG engine built around deep document understanding that turns complex files into a grounded, citation-backed question-answering layer. |
| Pathway | ★ 62.3k | A Python framework with a Rust streaming engine that keeps ETL, real-time analytics and RAG pipelines continuously up to date as source data changes. |
| Context7 | ★ 62.2k | Context7 pulls current, version-specific documentation and code examples for any library and feeds them into your LLM, available as a CLI skill or an MCP server. |
| LightRAG | ★ 39.7k | A graph-based RAG system that builds an entity-and-relationship knowledge graph for fast retrieval and easy incremental updates. |
| Quivr | ★ 39.5k | Quivr is an open-source RAG framework that ingests your documents and answers questions about them, working with any LLM and any file type. |
| Mike (MikeOSS) | ★ 4.3k | Self-hostable legal AI platform for document review, drafting and case-law research |