A local-first AI workspace — Chat, Documents, Sheets, and Benchmarking — running entirely on your machine. No cloud, no telemetry, no subscriptions.
Bring your own model: connect a local GGUF file, any OpenAI-compatible API (OpenAI, Ollama, LM Studio, etc.), or download a free model directly from the built-in gallery.
Built with Tauri v2 · React 19 / TypeScript · Python FastAPI · SQLite
🚧 Early Access — CrowForge is actively being developed. Things may break, features are still being added, and the project is evolving fast. Feedback, issues, and contributions are very welcome!
- AI Agent — ReAct-style agent that can autonomously read/write sheets and documents.
- Image Support — Insert images from disk directly into your documents.
- Document Search — Full-text search across your local knowledge base.
- AI Row Generation — Generate entire sets of sheet rows from a simple description.
- Data Privacy — One-click data deletion and backend management.
- Multi-session history with persistent local storage.
- Context modes: General, Writing, Coding, Analysis, Brainstorm.
- Document-aware Q&A: Attach your docs as context.
- Attach files (PDF, TXT, DOCX) directly to messages.
- Rich text editor with advanced AI writing tools (Rewrite, Summarise, Expand).
- AI Suggestions: Generate paragraphs, headings, and quotes in one click.
- Export to PDF, DOCX, or Markdown.
- Powerful local spreadsheet with unlimited rows/columns.
- AI Fill & Generate: Describe what you want, and the AI builds the data.
- Built-in templates for CRM, Task Lists, Budgets, and more.
- Export to XLSX or CSV.
- Compare multiple models side-by-side with the same prompt.
- Real-time latency, token count, and output comparison.
- Save and reload past benchmark sessions.
For detailed instructions on how to set up and build the project, or to understand its internal architecture, please refer to the following documents:
- 📥 Installation and Setup Guide — Prerequisites, environment config, and build steps.
- 🏗️ Architecture and Structure — Project layout and technical stack overview.
Made by Lubomir Timko (sanchez.sk), with AI assistance from Claude (Anthropic).
MIT — see LICENSE for details.