Mudroom lets founders design what buyers see when they research a product in AI. Add a small widget to your site, and visitors open ChatGPT, Claude or a coding agent with a guide you curate. See which guide pages their AI reads, and get the questions it can’t answer in Slack.
We draft your guide from your public site. You review it, pin suggested topics and add one script tag. Visitors choose their AI and bring along the guide, with an optional question. ChatGPT and Claude open with a prepared prompt; desktop visitors can also copy the prompt or connect Claude Code or Codex through MCP. Phones show ChatGPT and Claude.
Unanswered questions reach your team in Slack. Visitors receive replies on an answer page and by email.
A private stats page shows hand-offs by AI, fetched guide pages, topic-page reads and submitted questions. Hand-off-to-read matching is estimated; stats do not expose private AI conversations. During the pilot, Mudroom helps prepare your guide, suggested topics, branding and Slack hand-off. Book a 30-minute demo to walk through the visitor experience, setup and pilot pricing.
Read how it works for curating your guide, what visitors see, what you see, and how the pilot works.
For agents: /llms.txt summarizes this site, and /llms-full.txt describes Mudroom in one file. A guided version with suggested topics is at https://ai.mudroom.dev/llms.txt?via=llms; it ends with a link to ask our team. Coding agents can also ask us through our support door: https://support.mudroom.dev/llms.txt.
Mudroom also offers an open-source support door, where developers’ coding agents ask your team over HTTP and your team replies in Slack: mudroom-support on GitHub.