AI Agents
nanoclaw
A lightweight, container-isolated personal AI agent for WhatsApp, Telegram, Slack and more
★ 30.9k starsMIT
View guide →Fork() for AI agent microVMs — spawn 100 KVM-isolated children in ~100ms from a warm parent
forkd is an open-source ai agents project by deeplethe (deeplethe/forkd on GitHub). It has 2,923 GitHub stars and 227 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 14 September 2026.
Decide your fan-out shape: parallel agent evals, per-user code execution, untrusted CI, or fork-per-turn code interpretation.
Install the daemon (or build from source with Cargo) on a Linux host with KVM and Firecracker.
Run forkd doctor / quickstart to preflight KVM, Firecracker and networking, then pull or bake a warm parent snapshot.
Fork a few children from the snapshot tag, run a task in one, and check its timing and isolation.
The README reports spawning 100 children in about 100 ms from a warm parent snapshot, sharing memory copy-on-write until they diverge.
Pause a running sandbox, snapshot its in-flight state and resume — about 150 ms, with v0.4 live mode cutting the pause window further.
Each child is a separate Firecracker microVM with real KVM isolation, not just a container namespace.
Stack snapshot layers (e.g. Python base, then numpy, then pandas) that the daemon assembles in one pass at spawn time.
Pull portable prebaked snapshots such as a LangGraph parent image, or publish your own.
Python SDK (pip install forkd), TypeScript SDK (@deeplethe/forkd) and an MCP server (pip install forkd-mcp) for Claude and Cursor.
Clone forkd and enter the directory.
git clone https://github.com/deeplethe/forkd.git
cd forkdquickstart preflights the host (KVM, Firecracker, kernel image), bakes or pulls a Python snapshot, then forks 10 children with per-child timings.
sudo -E forkd quickstartPull a portable snapshot from the hub and fork children from it.
forkd pull deeplethe/langgraph-react
sudo -E forkd fork --tag langgraph -n 3 --per-child-netnsIt is maintained by deeplethe and contributors on GitHub.
Yes, it is Apache-2.0 licensed and self-hosted on your own Linux/KVM host, with no per-second cloud sandbox fees.
No — forkd is sandbox infrastructure, not an LLM service. Your agents inside the VMs use whatever model keys you give them.
Containers share a kernel; forkd gives each child its own microVM with KVM isolation, and unlike managed SaaS it runs on your own hardware.
On a KVM-capable Linux host, install forkd, run sudo -E forkd quickstart, then forkd pull a snapshot and fork your first children.
A lightweight, container-isolated personal AI agent for WhatsApp, Telegram, Slack and more
Orchestrate Claude Code, Kiro, and Codex together in isolated tmux sessions.
100+ AI Agents, Agent Skills and RAG Apps - Free and Open Source.
Governance and orchestration for AI agent fleets: declarative rules, verifiable runs.
The JavaScript/TypeScript version of LangChain - the agent engineering platform
AppAgent — Tencent's multimodal agent that learns smartphone apps by exploration, then operates them.
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if forkd is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.