Skip to main content

For your team

Your usability studies, where your team already works

Your developers and designers live in their terminal. Now they can pull what testers struggled with, check how a design scored, and draft a new study from there.

No dashboard detour. Connect once, and your OpenScouter workspace is a question away.

Works withClaude CodeOpenAI Codex

The toolkit

What your team can do from the terminal

See your studies and tests

List every study, where it stands, and which tester sessions have finished analysis.

Read the findings

Pull the barriers testers hit, ranked by severity, in plain English, with the WCAG reference attached.

Check the scores

Get a study or test score at a glance, including how testers experienced the design.

Draft a new study

Describe a page and the tasks you want covered, and your agent drafts it for you to review.

It can read and draft. It cannot spend.

Launching a study and paying testers is always a click in your dashboard. A connected agent can never spend money, recruit testers, or change a live study on your behalf.

You sign in once through your browser, and you can disconnect anytime in Settings, under Connections.

Scouty mascot with encouraging expression

One paste

Set it up in one paste

Copy the prompt below and give it to your Claude Code agent. It runs the steps for you. The only thing you do by hand is approve one browser sign-in.

paste into your coding agent
Set up OpenScouter in my terminal. OpenScouter is the usability testing platform my team uses to see how real people get on with our designs. I want to read our study findings and scores here, and draft new studies, without leaving my editor.

Please do the following:
1. Run /plugin marketplace add nicolaraimondo/openscouter-claude-plugin
2. Run /plugin install openscouter@openscouter
3. Run /mcp, pick openscouter, and give me the sign-in link to open in my browser so I can connect our workspace.
4. Once I am connected, list my studies so we can confirm it works.

If any step needs me to click or approve something, tell me exactly what to do.
Works withClaude CodeOpenAI Codex

Prefer to do it by hand?

In Claude Code, run these two commands, then type /mcp and pick openscouter to sign in.

/plugin marketplace add nicolaraimondo/openscouter-claude-plugin
/plugin install openscouter@openscouter

Using OpenAI Codex instead? Add a remote MCP server pointing at this URL, and Codex opens the same browser sign-in the first time you use it.

https://openscouter.com/api/mcp
Scouty mascot with thinking expression

Once you are connected, try asking

>List my studies and their status.
>Show me the findings for my checkout study.
>How did test 1234 score?
>Draft a usability study for our new pricing page.
Scouty mascot with celebrating expression

Not running studies with OpenScouter yet?

See how teams test their designs with real people before they ship, then bring it all into your terminal.