🩺 Vitals
- 🟢 Last active: 2026-08-05
- 📦 Latest release: v2026.7.1-2 (2026-08-04)
- 🐞 Open issues: 5511
- 🌟 Stars: 385.2k
What do these metrics mean?
- Last active: when code was last pushed, as of our last check. The dot is green when that was recent, grey otherwise. A long gap can mean a tool is finished and stable, not only unmaintained.
- Latest release: the most recent tagged, packaged version the maintainers published. Not every healthy project tags releases.
- Open issues: unresolved reports and requests. A high number is normal for a popular project and is not a warning on its own.
- Stars: how many people bookmarked the project on its forge. A rough popularity signal, not a measure of quality.
🏗️ Profile
- Official: openclaw.ai
- Source: github.com/openclaw/openclaw
- License: MIT
- Deployment: Desktop Binary | Docker
- Data Model: Local State / Plugin-driven
- Jurisdiction: Austria 🇦🇹 / EU 🇪🇺 (Individual Maintainer)
- Compliance (SaaS): N/A (No SaaS offering)
- Compliance (Self-Hosted): Self-Hosted (User Managed)
- Complexity: High (4/5) - Requires VM sandboxing for safe deployment
- Maintenance: Medium (3/5) - Rapid development cycle; high dependency churn
- Enterprise Ready: Low (1/5) - Not approved for general corporate use
1. The Executive Summary
What is it? OpenClaw (formerly known as Clawdbot and Moltbot) is an autonomous, open-source AI agent designed to execute tasks directly on a user's machine. Unlike traditional chatbots that merely generate text, OpenClaw can navigate the operating system, manipulate files, manage emails, and interact with web browsers on your behalf. While it offers unparalleled productivity gains, it represents a fundamental shift in the enterprise threat landscape by granting an LLM-driven process deep, persistent access to local resources.
The Strategic Verdict:
- 🔴 For General Employees: Hard Red Flag. The risks of prompt injection, malicious community-built "skills," and unintentional data deletion are catastrophic. Corporate workstations should strictly prohibit its installation.
- 🟢 For AI R&D Teams: Conditional Buy. For teams building the next generation of agentic workflows, OpenClaw is an essential experimental framework, provided it is deployed inside a strictly sandboxed, firewalled environment with zero access to production data.
2. The "Hidden" Costs (TCO Analysis)
| Cost Component | Claude Desktop (Proprietary) | OpenClaw (Self-Hosted) |
|---|---|---|
| Security Audit | Standard Review | Extensive (Sandboxing) |
| Data Leakage | Interaction Logging | High (Skill Vulnerabilities) |
| Governance | Venture-Backed | Chaotic (Community-Led) |
| Licensing | Per User/Month | $0 (MIT License) |
3. The "Day 2" Reality Check
🚀 Deployment & Operations
- Installation: Built as a natively-compiled agent that utilizes a plugin-based architecture for "skills." It requires local installation and thrives on high-performance machines with modern GPU acceleration for local model execution.
- Stability: Currently in an early, volatile state. The project has undergone multiple rebrands and architectural pivots in its first year, making it unsuitable for long-term production stability without internal dedicated maintenance.
🛡️ Security & Governance (Risk Assessment)
- Jurisdiction & Geopolitics (Austria 🇦🇹 / EU 🇪🇺): The individual maintainer is based in Austria, placing the project under EU jurisdiction by default. China has formally prohibited OpenClaw's use in state agencies, citing its deep OS privilege model as an automated data exfiltration risk. European enterprises benefit from EU GDPR jurisdiction but must independently assess whether OpenClaw's local execution architecture meets their data governance requirements.
- The Compliance Shift (Structural Security Gaps): OpenClaw ships with no RBAC, no centralized audit logging, and zero enterprise certifications. Safe enterprise deployment requires VM-level sandboxing, network isolation, strict secret management, and third-party security overlays, none of which are included in the core project. Organizations in regulated industries should treat this as experimental infrastructure, not production-grade software.
- License Risk (MIT: Operational, Not Legal): The MIT license is cleanly permissive with no copyleft clauses or enterprise tax. The risk is architectural: community-built skills from ClawHub have been found to contain supply chain threats including credential harvesters. The platform's deep OS access also makes it uniquely vulnerable to indirect prompt injection, where a malicious document can instruct the agent to execute unauthorized system commands entirely outside user intent.
4. Market Landscape
🏢 Proprietary Incumbents
- Claude Desktop: The market standard for integrated AI agents; users explore OpenClaw to regain control over the agent's logic and escape Anthropic's telemetry.
- MultiOn: A web-centric agentic service; OpenClaw provides a more raw, powerful, and local-first alternative for developers.
🤝 Open Source Ecosystem
- Accomplish: A more stable local agent that prioritizes secure browser and file workflows.
- Jan: A local-first AI orchestration desktop that prioritizes model portability and privacy-first interaction.