

Knowns CLI is an AI-first command-line tool designed for developers and teams to provide persistent project memory for AI assistants. It manages project context, tasks, and documentation, ensuring AI assistants always have the right context without repetitive re-explaining. The tool is built for real project handoffs, helping both solo developers and AI-assisted teams work efficiently with structured knowledge.
Without project memory, tasks, docs, and decisions are scattered across different tools, leading to lost context and repeated work. Every session starts with re-explaining the project, handoffs lose context, and AI assistants guess because they can't see the full picture. Knowns solves this by linking tasks to docs, specs, and decisions in one place, ensuring knowledge is extracted and reused instead of forgotten.
Knowns offers structured tasks with acceptance criteria, linked specs, and progress tracking. This allows users to plan work clearly and track completion against defined criteria, keeping everything organized and actionable within the project.
The tool includes capabilities for keeping project knowledge searchable, versioned, and close to work through its docs feature. This ensures that all project documentation is easily accessible and maintained alongside tasks, preventing knowledge from becoming orphaned or outdated.
Templates enable the reuse of proven workflows instead of starting from scratch each time. This saves time and ensures consistency across projects by leveraging past successful patterns and decisions.
Explicit references create links between tasks, docs, and decisions, ensuring nothing is orphaned. This interconnected system maintains context and makes it easy to navigate related project elements.
Semantic search is built in, allowing users to find anything by meaning, not just keywords. This enhances discoverability and helps teams quickly locate relevant information within the project memory.
Agent context gives AI assistants structured project memory via MCP integration. This allows AI tools to read and reference the full project context automatically, improving their assistance and reducing manual context sharing.
admin
Knowns works through a simple cycle: capture what needs to happen and what 'done' means, link tasks to docs and past decisions, work with AI reading the full context, verify acceptance criteria, and remember patterns for next time. This methodology ensures knowledge stays connected from idea to completion.
Benefits include eliminating repetitive context sharing, enabling AI to automatically reference project knowledge, and ensuring handoffs include history and linked references. Users experience no lost knowledge, reduced rework, and more efficient AI-assisted workflows.
Use cases include solo developers building projects across multiple sessions, where Knowns remembers architecture decisions and tracks progress. For AI-assisted teams, it ensures everyone works from the same structured context without conflicting assumptions. In product and engineering handoffs, specs, tasks, and acceptance criteria live in one system, providing full context for new tasks.
Target users are developers and teams, including solo developers, AI-assisted teams, and product and engineering teams. Integrations include MCP for AI agents and platforms like Claude. The tech stack is open source with a local-first approach, MIT licensed, and supports installation via Homebrew, shell script, PowerShell, npm, and npx. Pricing is not specified as it is open source.
In summary, Knowns CLI gives your project a memory by connecting tasks, docs, and decisions, ensuring AI assistants and teams always have the right context to work efficiently and avoid knowledge loss.
Knowns CLI targets developers and teams, including solo developers who build projects across multiple sessions and need memory for architecture decisions. It is for AI-assisted teams with multiple people and AI agents working on the same codebase to ensure structured context. Product and engineering teams benefit from handoffs where specs, tasks, and acceptance criteria are connected. Users value open source, local-first tools with no vendor lock-in, seeking efficient project memory for AI workflows.