Last updated: 2026-11-13
guilds.run runs teams of AI agents that you can inspect. Every prompt, tool call, memory change and agent-to-agent message is visible; you control permissions, edit context and choose the model for each agent, mixing Claude, GPT, Gemini and others in one team. It is open source and runs on your own machine. It launched on Product Hunt on 12 October 2026 and was ranked #4 of the day with 193 upvotes.
Visit guilds.run →Multi-agent workflows are easy to start and hard to trust. guilds.run's answer is transparency: when the system does something odd, you can see which agent did what and why, then adjust it. It ships ready-made "recipes" such as support triage and repository management.
Self-hosting is free. A hosted version was listed as coming soon on 13 November 2026.
You run and maintain it yourself, and model usage is billed by your model providers. Research published by Anthropic this week found that several cooperating agents often do worse than one coordinator, so visibility matters.
This profile is based on the maker's own pages and public launch records as of 13 November 2026, not on hands-on testing; capabilities and prices change often, so confirm on the vendor's site. Sources: guilds.run; Hunted.space: guilds.run launch data. AIHumanLove.com is independent: no ads, no affiliate links, no sponsored placements, and vendors cannot pay to change what we write. Compiled with AI assistance and reviewed before publication.