liliu-z/stashbase

An open-source, local-first writing workspace where you can write with Claude Code and Codex, drawing on your own sources and past work.

What it solves

StashBase provides a local-first writing workspace that functions like an IDE for writers. It solves the problem of fragmented writing workflows by integrating AI agents directly with a user's own local sources and past work, allowing for context-aware drafting and revision without losing the author's unique voice.

How it works

Users organize their work into local project folders. The system uses text extraction and semantic search to provide AI agents (such as Claude or Codex) with fast, accurate context from these files. It features a specialized "document diff" system that allows users to review AI-suggested edits as inline prose changes (red strikethroughs and green highlights) rather than traditional code-style diffs. It also supports the Model Context Protocol (MCP) to connect with external agents.

Who it’s for

It is designed for writers, researchers, and content creators who want to use AI to draft and refine documents while maintaining strict control over their local data and personal writing style.

Highlights

  • Local-first architecture: Documents stay in local folders and OCR/text extraction runs locally.
  • Semantic Search: Integrates with OpenAI or OpenRouter to search sources by meaning.
  • Prose-centric Diffing: Review AI edits directly in the flow of the document with intuitive Undo/Keep options.
  • Voice Preservation: Uses past writing and edits to guide new drafts.
  • MCP Support: Allows external MCP clients to read, search, and edit project files.

Related

  • Project
    swjybky/deepwrite

    A local-first AI writing workbench for creators that integrates specialized agents and a diff-based review system to collaboratively draft and refine manuscripts.

  • Project
    hAcKlyc/MyAgents

    MyAgents is an open‑source, local‑first desktop workbench for creating and orchestrating AI agents. It supports multiple LLM runtimes, a task scheduler, IM channel integration, and a private device‑to‑device network (AgentNet) for multi‑computer workflows, all under an AGPL‑v3 license.

  • Project
    busabase/busabase

    Busabase is an open‑source, local‑first database + workspace that lets AI agents (Claude Code, Codex, Cursor, Gemini, etc.) and people share structured records, docs, skills, and tiny apps. Every agent write is stored as a Change Request with diff, author, and history, enabling durable, auditable multi‑agent workflows. It runs via a simple `npx busabase server` command (or Docker/desktop), supports templates, and offers an API (MCP/OpenAPI) for any custom agent.

  • Project
    andrea9293/mcp-documentation-server

    A local-first document management and semantic search server for AI agents that provides a private knowledge base with a built-in web UI and hybrid search.

  • Project
    yilujian/easy-writing

    A privacy-focused desktop writing application for web-novel authors that combines a professional text editor with local-first AI assistance and project management tools.