Tale

Tale is an open-source workspace where human teams and AI agents work on projects together. Work is organized as tasks on a project board: you write the brief with source files and completion criteria, assign teammates or configured project agents, then follow progress, steer agents with @mentions, and review their reports and delivered files. A manager agent can keep the queue moving by delegating ready tasks, and repeatable processes can be captured as versioned automations with scheduled starts, workflow steps, and connector approvals.

Agents run in persistent sandbox workspaces with configurable concurrency, and each project agent can be set up with its own instructions, harness, model, and tools. Tale supports popular coding-agent runtimes including Claude Code, Codex, Cursor, Gemini CLI, OpenCode, and others, and works with your own provider API keys or supported vendor subscriptions. Built-in chat with side-by-side model comparison, shared project knowledge with retrieval, and governance guardrails round out the platform.

Tale is MIT-licensed and self-hostable with Docker Compose or the Tale CLI, with a managed cloud service and an Enterprise tier (same features plus professional operation and support) for teams that prefer not to run it themselves. Compared with proprietary project tools like Asana, Linear, or ClickUp, Tale adds the AI-agent execution layer natively; compared with agent platforms like Lindy AI or Relevance AI, it keeps the full project context — tasks, files, knowledge, and reviews — in one self-hostable workspace.