Kanwas

Kanwas is a context-first AI workspace that acts as a shared brain for product teams and agents. It combines a spatial canvas, a compounding knowledge graph, and model-agnostic AI agents that learn a team’s rules, workflows, and history. Instead of scattered chats, static docs, and forgotten whiteboards, Kanwas pulls code, tasks, research, conversations, and decisions into one Git-backed space where humans and agents reason over the same context. It targets product managers, founders, and cross-functional product teams that want sharper strategies, PRDs, and roadmaps without losing the human taste that makes those decisions distinct.

• Context-aware AI agents: Agents read boards, notes, tasks, and connected tools to draft strategies, specs, and research that reflect actual product history.
• Canvas for real work: Code, docs, tickets, embeds, and iframes live together on a spatial board so thinking, discussion, and AI automation happen in one place.
• Compounding context graph: Boards, notes, tasks, and decisions feed a graph that makes future recommendations and drafts progressively better aligned with prior choices.
• Model-agnostic stack: Teams can plug in Claude, GPT, Gemini, and other models while keeping the same workspace, prompts, and workflows.
• Git-backed markdown storage: Every document is a plain .md file with version history, supporting audits, portability, and local editing.
• Real-time collaboration: Multiple teammates and agents can work live on shared boards with fast sharing links and permission controls.
• 1,000+ connections and CLI: Integrations and a CLI bring context in from existing tools, making Kanwas a hub rather than yet another silo.

1. Sign up and create a workspace: Access Kanwas via the web app (hosted version) or deploy the open-source code on your own infrastructure.
2. Set up your first board: Start a spatial board and begin adding notes, documents, links, embeds, or iframes to map out a product decision, feature, or strategy.
3. Connect tools and integrate context: Use integrations or the CLI to pull in tasks, code, research, and conversations from your existing stack, so all relevant context lives together.
4. Configure AI agents and rules: Choose your preferred models (e.g., GPT, Claude, Gemini), and encode team rules, workflows, and decision-making patterns so agents generate drafts that reflect actual product history.
5. Collaborate and iterate: Invite teammates to the board, allow agents to propose specs or memos, review changes via Git-backed version history, and keep refining the workspace.
6. Let the graph compound: As boards, notes, and decisions accumulate, the knowledge graph improves future AI-generated drafts, creating a progressively smarter collaborative brain.

Kanwas follows a freemium model with a strong open-source foundation. The open-source edition is available on GitHub and is free in terms of software licensing—you handle your own hosting, infrastructure, and maintenance. A managed cloud service is accessible through the website, but specific per-seat or per-month pricing is not publicly listed; it is likely handled via early access or direct contact. For the most accurate and current pricing details, refer to the official Kanwas website.

• Open-source self-hosted: Free software (infrastructure and maintenance costs not included)
• Hosted workspace (cloud): Early access; contact Kanwas for pricing

• Web application: The primary interface available via the hosted cloud service or self-hosted deployment.
• CLI (Command Line Interface): For integrating context from other tools and automating workflows without leaving the terminal.
• Open-source codebase: Hosted on GitHub for self-hosting, inspection, and extension; the core software can run on your own infrastructure.
• Git-backed local editing: Since all content is markdown in a Git repository, you can edit documents locally using any text editor and push changes back.

• Commit to working inside Kanwas consistently: The compounding context benefits only appear when the team regularly uses the board as the primary thinking and decision space.
• Invest time upfront in wiring context: Connecting existing tools, encoding key rules, and importing historical decisions pays off as agents produce richer drafts.
• Use Git-backed version history as a safety net: Review diffs, roll back changes, and treat every markdown file as a portable artifact you can edit locally.
• Keep models interchangeable: Since Kanwas is model-agnostic, experiment with different LLMs for varying tasks without changing your workspace structure.
• Start with one high-value workflow: Rather than migrating everything at once, pick a critical process (e.g., PRD drafting) and expand into other areas as the graph grows.
• Leverage real-time collaboration: Use live boards with agents and humans simultaneously to minimize context switching and reduce meeting overhead.

• Product managers and product leads who want AI-generated specs and roadmaps rooted in real product history, not generic patterns.
• Startup founders and executive teams building a durable, shared context layer from customer interactions, strategy memos, and key decisions.
• UX researchers and designers who need to link research evidence to product boards so AI can surface relevant insights on demand.
• Engineering and platform teams looking to connect code constraints, incidents, and technical context directly to product priorities.
• Fractional product leaders working across multiple companies who need a portable, self-contained workspace with AI that adapts to each domain.
• AI agent builders and technical teams that value open-source code, markdown storage, and Git for customization and long-term data ownership.