Product updates, tutorials and industry insights to help you use AutoClaw better
AutoClaw V1.9.0 introduces GLM-5.2 with a 1M-token context window and Auto Design, a full-workflow design agent that turns complex PRDs into complete app and web interface designs.
Read More →
Cluster Mode enforces a strict SOP—plan, research, parallelize, audit, deliver—turning AI from "answering on instinct" to "delivering by process." Auditors catch real errors (PEG miscalculations, 71 ghost citations), parallel agents produce 20K-word reports in 43 minutes. Simple tasks solo, complex ones with a team.
Read More →
Run multiple independent AI agents on one machine—each with its own personality, memory, and workspace. Different agents, completely isolated memories; same agent across Discord/Telegram/WhatsApp, fully shared memory in real-time. Work assistant on Discord, life assistant on WhatsApp—two worlds, zero crossover. Privacy by design, no accidental leaks.
Read More →
Teach it once, it remembers forever. AutoClaw's Hermes self-evolution crystallizes your preferences, corrections, and workflows into persistent experience—across conversations, across restarts. Just say "From now on, keep it concise" and it proposes the rule; you approve, it sticks. It also self-learns from mistakes during complex tasks. Regular evolution captures single preferences; Skill evolution crystallizes entire workflows you can reuse with one sentence.
Read More →
Traditional AI chatbots can only answer questions and offer suggestions, leaving you to do all the actual work yourself. AutoClaw is different. Simply tell it "Help me organize the meeting minutes", and it will directly create a Feishu Doc, fill in structured content and send you the link. Say "Schedule a meeting with Luoran tomorrow afternoon", and it will check the calendar and send out meeting invitations automatically. It’s not just a chat tool — it gets things done.
Read More →