Building a better way to code with AI
Rekall was born from a simple frustration: existing AI coding assistants were either too generic, too limited, or required sending proprietary code to cloud services. We believed there had to be a better way.
What if an AI assistant could truly understand the context of your work? What if it could remember your preferences, learn from mistakes, and adapt to different development scenarios? What if you could keep everything local, private, and under your control?
Rekall is our answer to those questions. It's not just another AI chat interface — it's a comprehensive development partner with specialized modes, intelligent task management, and a self-improvement system that gets smarter over time.
The principles that guide Rekall's development
Your code is your intellectual property. Rekall works with local LLMs through Ollama, ensuring your code never leaves your machine unless you choose to use cloud models.
Rekall doesn't write placeholder code or TODOs. It generates complete, working implementations. If requirements are unclear, it asks questions rather than making assumptions.
Every interaction creates or modifies a task list. Work happens systematically, top to bottom, with progress tracking. No ad-hoc, unfocused coding sessions.
Rekall learns from every interaction. Problems are recorded, solutions are verified, and the system improves over time. The more you use it, the smarter it gets.
Rekall can create its own tools at runtime. If it encounters a task requiring a capability it doesn't have, it writes the tool, compiles it, and uses it immediately — extending its own abilities on the fly.
Different projects need different approaches. Task profiles provide specialized prompts and guardrails for web apps, libraries, games, and more.
A robust foundation for professional development
Built on Windows Presentation Foundation with .NET for a native Windows experience with professional performance and reliability.
Native integration with Ollama for running local LLMs. Use models like Llama, Mistral, or Qwen completely offline with no data leaving your machine.
Tools are C# classes that compile and load at runtime using the Roslyn compiler. Create new tools without restarting the application.
Task profiles are defined in JSON with markdown prompts and guardrails. Easy to read, easy to modify, easy to create new profiles.
Context, problems, solutions, and task history are persisted to JSON files. Your data survives restarts and is always accessible.
Comprehensive logging with Serilog for debugging and monitoring. Understand what Rekall is doing and troubleshoot issues easily.
The future of Rekall
Command-line interface for automation and IDE integration. Run Rekall from your terminal with the same powerful features as the GUI. Expose analyze, plan, and run phases for scripted workflows.
Model Context Protocol server for IDE integration. Connect Rekall to VS Code, Cursor, or other MCP-compatible editors. Expose tools and resources through a standardized protocol.
Additional task profiles for more languages and frameworks. Python, TypeScript, Rust, and more domain-specific configurations with specialized prompts and guardrails.
Smarter pattern recognition across problems and solutions. Proactive suggestions based on learned patterns. Better context understanding for more relevant assistance.
Active monthly and yearly subscribers receive all product updates.
Experience the future of AI-assisted development