Knowns CLI

A second core capability is its integrated documentation system with semantic search. Knowns maintains project knowledge in a versioned, searchable doc repository that stays close to the actual work. Unlike simple keyword search, Knowns includes built-in semantic search, allowing users and AI to find information by meaning and context, not just specific terms. This ensures that even vaguely remembered decisions or patterns can be retrieved efficiently. The documentation is not static; it is actively linked to tasks and decisions, creating a web of references that prevents knowledge from becoming orphaned. This feature group transforms scattered notes and specs into a living, interconnected knowledge base that evolves with the project and is instantly accessible.

Knowns further enhances productivity through templates and agent context via MCP integration. The templates feature allows teams to capture and reuse proven workflows and patterns instead of starting from scratch for every new task or project phase. For AI agent context, Knowns integrates with the Model Context Protocol (MCP), providing AI assistants with structured, real-time project memory. This means AI tools like Claude Code can read the full, organized context of a project directly through skills using the `/kn-*` namespace, such as `/kn-research`, `/kn-plan`, `/kn-implement`, and `/kn-extract`. These integrations enable AI to follow predefined plans, check acceptance criteria, and extract patterns for future use, effectively acting as an intelligent project participant with persistent memory.

The overall workflow of Knowns follows a simple, cyclical methodology designed to keep knowledge connected from idea to completion. It begins with the Capture phase, where users define what needs to happen and the criteria for completion. Next, in the Link phase, these tasks are connected to relevant docs, specs, templates, and past decisions. During the Work phase, AI assistants read this full, linked context to execute the plan. The Verify phase involves checking tasks against their acceptance criteria before marking them complete. Finally, the Remember phase involves extracting successful patterns and decisions into templates or documented knowledge for reuse in future projects. This cycle ensures context remains intact and continuously enriched, turning ad-hoc work into a repeatable, knowledge-preserving process.

Concrete use cases demonstrate how Knowns delivers tangible outcomes. For a solo developer building a project across multiple sessions, Knowns remembers architecture decisions, tracks progress, and gives their AI assistant the full picture every time they resume work, eliminating the need to reconstruct context. In an AI-assisted team with multiple human and AI agents working on the same codebase, Knowns ensures everyone operates from the same structured context, preventing conflicting assumptions and streamlining collaboration. For product-to-engineering handoffs, specs, tasks, and acceptance criteria live in one connected system; when engineering picks up a task, the full context—including linked docs and related decisions—is already present, clarifying the definition of done and accelerating implementation, as shown in the demo where skills implement dark mode in a Flutter project.

Knowns is built for specific target users including solo developers, AI-assisted teams, and product/engineering pairs who need to maintain project context across sessions and handoffs. It is a command-line tool compatible with macOS, Linux, and Windows, installable via Homebrew, shell script, PowerShell, npm, or npx, requiring Node.js 20+ for npm-based installs. The tech stack supports a local-first architecture with a web UI for kanban, docs, and chat, and integrates with AI platforms like Claude. As an open-source tool under the MIT license, it offers transparency and avoids vendor lock-in. The primary takeaway is that Knowns transforms project knowledge from a fragmented, ephemeral burden into a structured, persistent asset that empowers both human teams and AI assistants to work more efficiently and intelligently.

Updated 2026-02-28