kirodotdev/KiroCrew

A persistent workspace for development work that self-improves and continues beyond one session.

What it solves

Kiro Crew provides a persistent, self-learning development workspace that allows AI agents to continue working on tasks unattended, across multiple sessions, and on hardware controlled by the user. It eliminates the problem of ephemeral agent sessions that end when a chat window is closed.

How it works

The system uses a "Gateway" architecture that separates the agent runtime from the user interface. The Gateway manages sessions, memory, schedules, and approvals on the host machine, while the user interacts with the agent via a desktop app, web dashboard, CLI, or integrated messaging platforms like Slack and Discord. It utilizes an ACP runtime and supports MCP tools for extensibility.

Who it’s for

Developers who need autonomous, long-running AI agents that can remember project context, learn from failures, and execute complex multi-step tasks without constant supervision.

Highlights

  • Persistent Memory: Sessions and task checkpoints survive restarts, allowing work to resume exactly where it left off.
  • Self-Learning & Evolving: Converts corrections and failures into durable lessons and synthesizes repeated patterns into reusable skills.
  • Unattended Autonomy: Supports scheduled recurring jobs and reactive work triggered by webhooks or messaging events.
  • Multi-Surface Access: A single agent state can be accessed across a desktop app, web dashboard, CLI, and various chat apps.
  • Parallel Execution: Ability to spawn isolated subagents to handle parallel research or tasks before synthesizing results.
  • Defense in Depth: Includes OS sandboxing, tool approvals, and credential redaction for secure local execution.

Related

  • Project
    mikeyobrien/rho

    An always-on personal AI operator with persistent local memory and proactive scheduled check-ins, accessible via CLI, Web UI, Telegram, and email.

  • Project
    kirodotdev/Kiro

    An agentic AI development platform that helps developers plan, build, and review software by turning prompts into executable specs and using parallel agents across multiple interfaces.

  • Project
    SII-Holos/synergy

    Synergy is an open‑source, local‑first platform that makes AI‑agent work durable and recoverable across web, desktop, CLI, and API surfaces. It lets agents keep history, files, browser sessions, and tool configurations in a single workspace, supports specialist sub‑agents and reusable Blueprints, and offers extensibility via plugins. Benchmarks show its coding agent (Forge) improves DeepSWE Pass@1 to 67.3% at $0.54/task. Install via desktop installers or a one‑line CLI script, then configure a model provider and start the runtime.

  • Project
    peterfei/ai-agent-team

    A virtual AI development team with specialized roles (PM, Backend, Frontend, etc.) and a persistent memory system for managing complex software projects.

  • Project
    elie222/rakazo

    Rakazo is an open-source, self-hosted platform for running persistent AI agents that can access local files, browsers, and terminals across desktop and mobile devices.