Independent register of AI agents & harnesses · no sponsored placementsEdition 2026-09 · 107 entries · evidence to 26 Sept 2026

Compare / OpenClaw vs ZeroClaw

OpenClaw vs ZeroClaw

OpenClaw rates 63 against ZeroClaw's 61 in this edition. The table lines up every dimension we record; the leading value in each numeric row is marked. Add a third or fourth entry.

Dimension
OpenClawOpenClaw Foundation (created by Peter Steinberger)
ZeroClawZeroClaw Labs
Rating
AgentsWisdom score6361
Confidencehighmedium
Rank in categoryNo. 2 of 3No. 3 of 3
Pillars
Adoption & momentum9871
Experts (independent reviews)5967
Crowd (community sample)3747
Community posts sampled259
Experts vs crowdCrowd cooler by 22Crowd cooler by 20
Adoption
GitHub stars391k33k
npm weekly downloads2.8Mn/e
PyPI monthly downloadsn/en/e
VS Code installsn/en/e
Reported usersn/en/e
Reported customer organisationsn/en/e
Releases, last 90 days313
Benchmarks (shown, not scored)
Terminal-Bench 2.0not listednot listed
SWE-benchnot listednot listed
Independent reviews83
Trust & safety
Overall gradeBB
Permission modelCA
Data access scopeBA
Data storageAA
Data retention & trainingn/en/e
Incident historyCC
ComplianceDD
TransparencyAA
Price
PriceFreeFree
Plans
  • Open source (self-hosted): Free
  • Open source (self-hosted): Free
Product
CategoryPersonal & desktop agentsPersonal & desktop agents
MakerOpenClaw Foundation (created by Peter Steinberger)ZeroClaw Labs
LicenceMITMIT OR Apache-2.0
Use it viaCLI, Desktop app, Web app, Mobile app, Chat appsCLI, Chat apps, Web app
Runs onmacos, linux, windowslinux, macos, windows
Models it drivesClaude, GPT, Gemini, Grok, DeepSeek, local via Ollama, any OpenAI-compatibleClaude, GPT, local via Ollama / LM Studio / llama.cpp / vLLM, any OpenAI-compatible
Key features
  • Local Gateway process that binds to loopback by default
  • 29 messaging channels including WhatsApp, Telegram, Discord, Slack, iMessage and Signal
  • Persistent memory, skills and plugins, with the ClawHub skill registry
  • Shell exec, file operations, browser control and cron-style scheduled tasks
  • Optional Docker/Podman/OpenShell tool sandbox and exec-approval policies
  • Shared cloud sessions and rebuilt browser Control UI in 2.0 (v2026.8.1)
  • Can use existing ChatGPT/Claude subscriptions, API keys or local models
  • Single static Rust binary; targets low-memory hardware such as Raspberry Pi
  • 30+ channels including Telegram, Discord, WhatsApp, Slack, Signal, Matrix and email
  • Default 'supervised' autonomy: low-risk runs, medium-risk asks, high-risk blocked
  • Auto-detected OS sandbox (Landlock/Bubblewrap/Firejail, Seatbelt, AppContainer, Docker)
  • Workspace-only file and shell boundaries with forbidden paths
  • HMAC tool receipts and an emergency stop command
  • Gateway with web dashboard for chat, memory, config and cron
  • SOP engine, WebAssembly plugin host, A2A discovery, ZeroRelay/ZeroRouter (v0.8.x)
Verdicts
Praised for
  • Broad reach into everyday tools; users report automating real business workflows (HN, Reddit)
  • Large, fast-moving open-source project with hundreds of contributors per release
  • Highly customisable: multiple agents, per-project threads, local or hosted models
  • Small resource footprint compared with OpenClaw (independent reviews, community reports)
  • Security-by-default posture: supervised autonomy, sandboxing and workspace scoping
  • Some users report it as stable and easy to run on a small VPS
Criticised for
  • Security exposure: repeated advisories, a 2026 malicious-skill campaign, and a trusted-operator default with sandboxing off
  • Upgrades frequently break installs; users report hours of repair work after 2.0 and 2026.7.x updates
  • Setup and maintenance are demanding for non-technical users
  • Newer safety limits removed capabilities some users relied on
  • Several users report bugs, tool-calling failures and freezes, especially with local models
  • Smaller plugin/integration ecosystem than OpenClaw; Rust makes extension harder for some
  • Limited visibility into running tools in early versions (HN)
  • Impersonation sites and forks caused confusion; the official repo was briefly unavailable (404) in late March 2026
Evidence
Evidence collected26 Sept 202626 Sept 2026
Sources cited64