A beta program generates a steady stream of things that need reading. Tickets arrive overnight. Step results pile up. Testers who were active in week one go quiet in week two, and nobody notices until the program is over.
Until now, all of that lived behind the Stomio manager view, which meant it needed you to go there and look.
From today it does not have to. Stomio is a remote MCP server, so the AI assistant you already have open can work on your beta program directly.
What does that mean in practice?
MCP is an open standard for connecting an AI assistant to a system that holds real data. Stomio publishes more than eighty tools through it, covering beta projects, tasks and steps, tickets, testers, announcements and applicant screening.
The difference shows up the first time you ask a question. Before, an assistant could tell you how beta programs usually work. Now it can tell you how yours is going, because it is reading your tickets and your results while it answers.
It is the same platform, with the same permissions and the same rules. You are just reaching it from a conversation instead of a browser tab.
How do you connect it?
Roughly two minutes, and no engineer, API key or installation is involved. Here is the whole thing on Claude:
The short version: copy your connect URL from Workspace settings, then AI tools. Paste it into your assistant wherever it adds a connector or a remote MCP server. Sign in to Stomio, choose what the connection is allowed to do, and approve it.
It is the same URL for every assistant, so claude.ai, Claude Desktop, Claude Code, ChatGPT and Cursor all connect the same way. The step-by-step version, including the client settings a few assistants ask about before they will connect, is in Connect your AI assistant to Stomio.
What is worth asking it first?
The questions that tend to land best are the ones that were tedious to answer by hand:
- "Summarize what my testers reported this week and tell me what to fix first."
- "Which tasks are testers dropping out of, and what do they say when they do?"
- "Find every tester on firmware 2.1 who reported a connection problem."
- "Triage the tickets that came in overnight and tell me which ones block the release."
- "Draft the week-two task set based on the program we ran on the last version."
- "Write the mid-beta announcement, mention the two fixes that shipped, and send it to the testers who finished week one."
The last two are the interesting ones. They are the point where an assistant stops describing your beta and starts doing the work inside it.
What will it never do on its own?
Giving an assistant access to a live beta program only works if the limits are real, so we built them before we built the tools.
You choose the access when you connect. Reading comes with the connection. Everything that writes is a separate approval on the consent screen, granted as a whole or not at all: announcements, testers and applicants, tickets, tasks and steps, and beta projects. A connection approved for reading only cannot write anything, whatever it is asked to do. Which of the write tools your assistant reaches for after that, and when it checks with you, is your assistant's own setting.
Anything that reaches a person is previewed first. Announcements, invitations and ticket replies come back as a preview showing who they would reach and what they would say, and Stomio does nothing until that preview is confirmed. In a conversation you are having, the confirmation is yours to give. An assistant configured to approve its own tool calls can confirm without stopping to ask, so grant write access with the way you run that assistant in mind.
It only sees what you see. The connection carries your own permissions, never more, and it only reaches beta projects that have AI features turned on. A beta with that switch off is invisible to every assistant.
Contact details do not leave Stomio. Email addresses, phone numbers and precise locations are stripped from results, except where an email address is the only identifier a person has, which is the case for applicants and for people whose invitation is still pending.
You can end it in one click. Every connected assistant is listed under Workspace settings, then AI tools, with what it was allowed to do and when it was connected. Disconnecting takes effect immediately, per assistant.
One more thing that comes up often: this does not touch your Stomio AI credits. Those cover the AI features inside Stomio, such as AI Canvas and the AI insights buttons. Work you do through your own assistant runs on that assistant's subscription.
Where to start
If you already use an assistant daily, connect it read-only and ask it what your testers said this week. That one question is usually enough to tell whether the rest of it is worth your time.
You can read more about how AI works across Stomio on our AI page, or go straight to the connection guide.
