Key takeaways
- 15+ chat platform adapters — Telegram, Discord, Slack, WhatsApp, iMessage, Matrix, Feishu, Signal, DingTalk, QQ, IRC, Email, Nostr, and Web
- Full agentic loop with up to 100 tool iterations per request — shell, files, web browsing, search, scheduling, and persistent memory
- Built in Rust with NanoClaw-inspired design — channel-agnostic core with platform adapters, Docker sandbox support
- Plugin system, MCP support, plan-and-execute mode, and semantic memory with optional sqlite-vec embeddings
- Doubled stars since February 2026 (358 to 715) with rapid release cadence — v0.2.2 shipped June 3, 2026, plus 42 built-in skills and concurrent specialist sub-agents
FAQ
What is MicroClaw?
MicroClaw is a Rust-based agentic AI assistant for chat surfaces. It connects to 15+ messaging platforms and executes multi-step tasks using shell commands, file operations, web browsing, and more.
How does MicroClaw differ from NanoClaw?
MicroClaw is inspired by NanoClaw but built in Rust instead of TypeScript, with broader platform support (15+ vs NanoClaw's WhatsApp focus) and semantic memory via sqlite-vec.
How much does MicroClaw cost?
Free and open-source. Requires an LLM API key (Anthropic or OpenAI-compatible).
Who competes with MicroClaw?
NanoClaw (TypeScript, container isolation), Moltis (Rust, single-binary gateway), ZeroClaw (Rust, security-first), and OpenClaw (TypeScript, maximum features).
Executive Summary
MicroClaw is a Rust-based agentic AI assistant that lives in your chats. Inspired by NanoClaw, it takes the channel-agnostic approach further with 15+ platform adapters, full tool execution (up to 100 iterations per request), and optional semantic memory via sqlite-vec embeddings. [1] As of June 2026 it is in active development — stars have doubled to 715 since this profile's first publication, v0.2.2 shipped June 3, 2026, and commits land near-daily. [2]
| Attribute | Value |
|---|---|
| Author | everettjf (microclaw org) |
| Language | Rust |
| License | MIT |
| GitHub Stars | 715 (as of June 2026) |
Product Overview
MicroClaw uses a channel-agnostic core with platform adapters, supporting Telegram, Discord, Slack, Feishu/Lark, Matrix, WhatsApp, iMessage, Email, Nostr, Signal, DingTalk, QQ, IRC, and Web. Every message triggers an agentic loop where the model can call tools, inspect results, and reason through multi-step tasks. [1]
Since February 2026 the project has repositioned as a "Rust Agent Runtime": the v0.2.x line added 42 built-in skills with ClawHub installs, concurrent specialist sub-agents, MCP tool federation, and a reflector-based memory lifecycle (extraction and dedupe) on top of file plus structured SQLite memory. [3]
Key Capabilities
| Capability | Description |
|---|---|
| 15+ Platforms | Telegram, Discord, Slack, WhatsApp, iMessage, Matrix, Signal, Feishu, DingTalk, QQ, IRC, Email, Nostr, Web |
| Agentic Loop | Up to 100 tool iterations per request |
| Tool Execution | Shell, files, web browsing, search, code editing |
| Persistent Memory | Cross-conversation with optional semantic search (sqlite-vec) |
| Scheduling | Timezone-aware task scheduling |
| Skills | 42 built-in skills plus ClawHub skill installs |
| Sub-Agents | Concurrent specialist sub-agents (researcher, coder, writer, analyst) |
| Plugins | Extensible plugin system |
| MCP Support | Model Context Protocol tool federation |
| Plan & Execute | Multi-step task planning mode |
| Docker Sandbox | Isolated execution environment |
| Diagnostics | microclaw doctor for preflight checks |
Installation
One-line install or Homebrew on macOS: [1]
curl -fsSL https://microclaw.ai/install.sh | bash
# or
brew tap microclaw/tap && brew install microclaw
Strengths
- Broadest platform coverage — 15+ chat platforms, more than most competitors
- Rust performance — Single compiled binary, memory-safe, fast startup
- NanoClaw-inspired design — Channel-agnostic core proven by NanoClaw's 10K stars
- Semantic memory — Optional sqlite-vec embeddings for intelligent recall
- Full agentic capabilities — Shell, files, web, search, scheduling in one package
- Cross-platform — macOS, Linux, Windows with dedicated installers
- MCP support — Extensible via Model Context Protocol
Cautions
- Early stage by the author's own admission — at launch the creator called it "still a toy / experimental project" with no sandboxing or security hardening; the project has matured since, but treat shell-execution defaults carefully [4]
- Rapid churn — versioning jumped from v0.1.x to v0.2.x within weeks; features and config may change between releases [2]
- Rust contribution barrier — Higher than TypeScript/Python for casual contributors
- NanoClaw derivative — Inspired by NanoClaw; differentiation still emerging
- Single-maintainer concentration — everettjf has authored roughly 890 of the project's commits; bus factor is effectively one
- Smaller ecosystem — Fewer plugins and community resources than OpenClaw or NanoBot
What Developers Say
Independent third-party reviews remain scarce as of June 2026 — most public commentary comes from the maintainer or aggregator listings, which is itself a signal of how early this project is.
The creator (everettjf) was candid at the Hacker News launch in February 2026: [4]
"this is still a toy / experimental project"
He described it as "vibe-built" in about a day as a learning exercise and warned against running it on systems you care about without understanding the risks. [4]
The Shelldex registry, which tracks OpenClaw-ecosystem projects, rates MicroClaw at a 75/100 health score with "active development" status and ranks it #31 on its leaderboard. [5]
Pricing & Licensing
| Tier | Price | Includes |
|---|---|---|
| Open Source | Free | Full assistant, MIT license |
No paid tier, hosted offering, or public funding has been disclosed. [3]
Hidden costs: LLM API usage (Anthropic or OpenAI-compatible).
Competitive Positioning
| Competitor | Differentiation |
|---|---|
| NanoClaw | NanoClaw focuses on container isolation with WhatsApp; MicroClaw has 15+ platforms in Rust |
| Moltis | Moltis is a Rust gateway/router; MicroClaw is a full agentic assistant |
| ZeroClaw | ZeroClaw prioritizes security (WASM sandbox); MicroClaw prioritizes platform breadth |
| OpenClaw | OpenClaw has 50+ channels but 430K LOC; MicroClaw aims for lighter Rust alternative |
Bottom Line
MicroClaw fills a clear niche: a Rust-based personal AI assistant with the broadest chat platform coverage in its weight class. It combines NanoClaw's channel-agnostic philosophy with Rust's performance and 15+ platform adapters. Four months after launch the trajectory is encouraging — stars doubled to 715, near-daily commits, and a v0.2.x line with skills and sub-agents — but it remains a young, effectively single-maintainer project that began life as a self-described experiment.
Recommended for: Users wanting a single agentic assistant across many chat platforms with Rust performance.
Not recommended for: Users needing mature ecosystem, extensive plugins, or minimal setup (try OpenClaw or Lindy).
Research by Ry Walker Research • methodology