Index / ZeptoClaw · updated Sep 4, 2026

ZeptoClaw

qhkm/zeptoclaw · healthy · rank 29 of 43 by stars

A single ~6MB Rust binary that packs tools, memory, channels, 18 providers, and sandboxed autonomy into a 50ms-startup personal AI assistant. It explicitly positions itself as the synthesis of OpenClaw's ecosystem, NanoClaw's isolation, NemoClaw's guardrails, and PicoClaw's size discipline.

#rust#local-first#single-binary#sandboxed#edge
Compare vs OpenClaw
GitHub ↗

Facts

Repository

Contributors16
Open issues53
Last commitJul 24, 2026
Release cadence~3 days
Latest releasev0.9.2

Runtime

LanguageRust
Memory6 MB
Boot time50 ms
Deploymentedge · desktop · self-hosted · cloud
Setup difficultyLow
Plugin ecosystemEmerging

Posture

LicenseApache-2.0
Local-firstYes
Cloud dependencyOptional
Privacy postureStrong

Community

Sentiment55% positive
Reddit mentions0
Web results6

Security breakdown

Composite 82 / 100 · how these are scored

Sandboxing8 / 10

higher is safer

API security7 / 10

higher is safer

Network isolation7 / 10

higher is safer

Telemetry safety8 / 10

higher is safer

Shell access risk4 / 10

higher is riskier

Model access

Read from the repository, not written by a model · 54 files examined

5 providers · gateway support

Direct
AnthropicGoogleGroqOpenAIVertex AI
Gateway
OpenRouter
Compatible
OpenAI-compatible

Pinned models

gpt-4oOpenAI
gpt-5.4OpenAI

gpt-4o released 2024-05-13 — 2 years old (from the public model catalogue)

Pin last edited 5 months ago

Several defaults are pinned; the project picks one per provider rather than shipping a single default.

Evidence

Decision

Why choose ZeptoClaw over OpenClaw?

Why choose this

  • ~6MB binary and ~6MB RAM vs OpenClaw's large TypeScript app
  • ~50ms startup, enforced by a CI binary-size gate
  • Built-in sandboxed autonomy and guardrails without Docker/k3s

Tradeoffs

  • Far smaller community and ecosystem (648 stars, no Reddit presence)
  • No comparable skills registry or third-party integrations depth
  • Younger project with less production battle-testing

Best fit

  • Users wanting OpenClaw-style features without the TypeScript footprint
  • Edge, VPS, or low-resource deployments
  • Rust developers who value a single auditable binary

Avoid if

  • You need OpenClaw's mature skills/plugin ecosystem
  • You want a large community and extensive third-party docs
  • You need proven multi-user team deployments
Good Confidence72%

Strong evidence from the README and active, disciplined commit history (CI gates, security fixes, dependency hygiene). However, there is zero Reddit discussion and minimal independent web coverage, so community sentiment and real-world adoption claims are uncertain.

AI layer reviewed Aug 24, 2026 · how this is written

Star activity

649 stars today

Overview

ZeptoClaw is a local-first personal AI assistant infrastructure written in Rust, shipped as a single ~6MB binary that starts in ~50ms and idles at ~6MB of RAM. It bundles tools, long-term memory, messaging channels (including a streaming Telegram gateway), 18 LLM providers, and sandboxed autonomy into one deployable unit suitable for desktops, VPSs, and edge hardware.

The project explicitly frames itself as a synthesis of its predecessors: OpenClaw's broad integrations without the TypeScript footprint, NanoClaw's container-isolated simplicity, NemoClaw's policy-gated guardrails without Docker/k3s, and PicoClaw's edge efficiency with Rust's safety guarantees. Engineering discipline is a visible priority — CI enforces a 7.5MB binary-size ceiling on every PR to protect the 'fits on a robot' moat, and the repo carries 3,900+ tests plus active dependency and security maintenance (cargo-deny, bcrypt/mail-parser bumps, provider-resolution security fixes).

Compared to OpenClaw, ZeptoClaw trades ecosystem breadth for footprint, startup speed, and a hardened single-binary deployment story. It is a strong fit for self-hosters and edge deployments who want assistant infrastructure they can audit and run anywhere, but users who depend on OpenClaw's skills registry and community momentum will find it less mature.

Nominate a clone

Add a new Claw

Paste a GitHub repository and tell us why it belongs on the tracker.

Opens a prefilled issue on GitHub — every nomination is public. Comfortable with a PR? Adding the repo to projects.json is faster.