Knowns: The Local-First Context Layer for AI-Native Development

Knowns is a tool designed to help developers work better with AI coding assistants. It stores project information like tasks, documentation, and architectural decisions in a special folder called.knowns. This setup allows AI tools like Claude, Cursor, or Copilot to read the project's history and rules without needing constant explanations from the human developer. Instead of starting from zero every time, the AI can pick up where previous sessions left off and build code with full awareness of the project's context.

Benefits

Knowns solves the problem of stateless AI sessions where every chat starts fresh. This leads to wasted time re-explaining architecture or pasting documentation into chat windows. Knowns fixes this by providing persistent, structured access to project knowledge. It helps developers save time by eliminating the need to repeatedly clarify past decisions or rebuild context that took hours to set up. The tool uses local storage, meaning it runs offline and does not require API keys. It also includes a three-layer memory system that ensures the AI recalls project decisions, patterns, and conventions without repetition. Developers can search by meaning rather than just keywords using semantic search powered by ONNX models. The system supports code intelligence through AST-based indexing for languages like Go, TypeScript, JavaScript, and Python. This allows the AI to search symbols, trace dependencies, and explore the codebase structure efficiently. A built-in web interface provides a Kanban board, document browser, and knowledge graph visualization for better project management.

Use Cases

Knowns is ideal for development teams who pair with AI coding assistants daily. It works well for managing tasks with acceptance criteria and implementation plans. Teams can use it to organize project knowledge in nested markdown folders with cross-references. The tool is useful for creating templates to generate code consistently across the project. It supports AI agent workspaces that handle multi-phase tasks like research, planning, implementing, and reviewing code automatically. Developers can use the web UI to visualize the project structure and track progress on tasks. The tool is also helpful for importing documentation and templates from Git repositories or local sources. It is suitable for both individual developers and larger teams who want to maintain full control over their data without relying on cloud services.

Pricing

Pricing details are not available in the provided information.

Vibes

Public reviews or testimonials are not available in the provided information.

Additional Information

Knowns is built using Go and requires version 1.24.2 or higher. It supports installation via Homebrew, shell scripts, PowerShell, npm, or by building from source. The project is open source and hosted on GitHub under the knowns-dev organization. The current version is 0.18.0. The roadmap includes active work on AI Agent Workspaces with git worktree isolation and live terminal streaming. A planned feature is a self-hosted team sync server to allow shared visibility without giving up the local-first workflow. The tool is designed to run fully offline and requires no API keys for its core functionality.