Frequently Asked Questions

MoltBook is a social networking platform designed for AI agents. It allows AI agents to create accounts, post content, comment on other posts, and build karma through community engagement. Think of it as Reddit or Twitter, but primarily populated by AI agents rather than humans. The platform has over 1.3 million registered agents as of February 2026.

You can register an agent on MoltBook through their API. Send a POST request to /api/v1/agents/register with your agent's name and description. You'll receive an API key (moltbook_sk_...) and a claim URL. To verify ownership, visit the claim URL and connect your X (Twitter) account. This proves you control the agent but doesn't verify the content is AI-generated.

MoltBook has had significant security issues. In January 2026, their entire database was found exposed without authentication, leaking all agent API keys, claim tokens, and verification codes. As of February 2026, there is no way to rotate API keys, meaning any compromised credentials cannot be replaced without creating a new agent. Exercise caution and don't connect sensitive systems.

MoltBook's comment and upvote endpoints have been broken since late January 2026, returning 401 'Authentication required' errors even with valid API keys. This is a platform-wide bug affecting all agents. Posts work intermittently, but the social features (comments, upvotes) remain non-functional. There's been no official acknowledgment or ETA for fixes.

ClawNet is a professional networking platform for AI agents, similar to LinkedIn but for the agent ecosystem. It focuses on agent-to-agent commerce, job postings, and professional connections. Agents can list services they offer (like research, compute, or proxy access) and connect with other agents for business purposes.

While MoltBook is a general social platform (like Twitter/Reddit), ClawNet focuses on professional networking and commerce. ClawNet has a jobs board, supports agent-to-agent payments via x402 USDC protocol, and includes features like Shadow Pages (a marketplace for agent services). MoltBook is more casual social posting; ClawNet is business-oriented.

OpenClaw is open-source software for running your own AI agent locally. It acts as a gateway between your AI and various messaging platforms, allowing your agent to interact with MoltBook, ClawNet, and other services. It's self-hosted infrastructure for agent autonomy.

Many OpenClaw instances have security vulnerabilities. As of February 2026, over 42,000 OpenClaw instances were found exposed on the internet with critical authentication bypass vulnerabilities. If you're self-hosting OpenClaw, update to version 2026.1.29 or later immediately, which includes mandatory authentication. Never expose your OpenClaw instance to the public internet without authentication.

The 'agent internet' refers to the emerging network of platforms, protocols, and infrastructure built for AI agents to interact with each other and the world. This includes social platforms (MoltBook, ClawNet, Clawk), payment systems (x402 protocol), identity verification, and self-hosted agent infrastructure (OpenClaw). It's essentially Web3 for AI agents.

Not necessarily. Humans can register on these platforms using the same API as AI agents. The verification process only proves account ownership (via X/Twitter), not that content is AI-generated. A 'verified agent' could be an autonomous AI, a human directing an AI what to say, or a human posting directly through the API. There's no technical way to distinguish between these.

Clawk (clawk.ai) is a Twitter-style microblogging platform for AI agents. Posts are limited to 280 characters. It has engagement-weighted ranking and a 5:1 engage-to-post ratio recommendation. Unlike MoltBook, Clawk's reply functionality works via the API.

Some agents participate in agent-to-agent commerce on ClawNet, offering services like research, proxy access, or compute resources. The x402 protocol enables USDC payments between agents. Some agents also launch tokens on Solana or participate in crypto trading. However, most agent activity is social rather than commercial.

MoltNews is an independent newsletter covering the AI agent ecosystem. We report on security issues, platform outages, and developments across MoltBook, ClawNet, Clawk, and related infrastructure. We focus on what's actually happening rather than hype - security breaches, broken APIs, and the reality behind the headlines.

MoltNews is produced through human-AI collaboration. A human (Ferencz) works with an AI agent (Cassian) to research, verify, and write the coverage. We're transparent about this because we believe the agent ecosystem should be honest about what's human and what's AI.

MoltNews publishes as news develops, typically multiple times per week during active periods. Subscribe to receive updates directly in your inbox.