Mimiq: AI agents use your app like your users and find what breaks

Working · 1.5 s

io.github.victorgulchenko/mimiq-mcpv0.5.1

AI agents use your app like your users, in real browsers, and report what is broken or confusing.

Developer Tools Browser Automation

Source Website Registry entry

Maintain this server? Claim it

Try it

Call this server's tools from your browser. Nothing to install.

https://mcp.mimiqai.com/mcp

Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.

Tool list from our last check. Connect live to refresh it.

mimiq.show_report Read-only

Show the finished result of a saved Mimiq run: what the agents hit, where each stopped and why, and a link to the full report. Use the session ID from an earlier run or a report URL. Reads your account only and spends no credits.

session_idstring, required

compare_session_idstring

Tools (2)

Tool list last changed Oct 23, 2026.

mimiq.show_report

read-only

Argument	Type	Description
session_id*	string
compare_session_id	string

Input schema (JSON)

{
  "type": "object",
  "required": [
    "session_id"
  ],
  "properties": {
    "session_id": {
      "type": "string",
      "pattern": "^[A-Za-z0-9_-]{1,100}$"
    },
    "compare_session_id": {
      "type": "string",
      "pattern": "^[A-Za-z0-9_-]{1,100}$"
    }
  },
  "additionalProperties": false
}

mimiq.test_flow

Use this after you build or change anything a user goes through (signup, onboarding, checkout, settings, a new feature), on the preview or staging URL, and again after each fix. AI agents, each acting as one of the app's users, open it in their own real browser and try the task by clicking, typing and filling in forms (they get a working inbox for verification codes). They report what was broken, what did nothing, where they got stuck or gave up and why, with the step and a screenshot, and the problems ranked by how many agents hit them. Pass the URL and, as goal, the task a user would try in plain words. For a protected preview or pages behind a sign-in, pass access (a Vercel bypass secret, a username and password, or a test account). Each agent keeps going until it finishes the task or gets stuck, so a run takes a few minutes, longer for a long task; each agent uses 5 credits. It starts right away and the result includes the audience summary, examples and fit. If the recruited crowd is not the requested audience (more than 1 in 5 wrong) or the page could not be read to check its market, the result is an audience_review instead: show the warning, then continue with review_id, confirm_audience and audience_fit_acknowledged, or recruit again with audience_correction.

Argument	Type	Description
url*	string	Where the flow starts. Must be publicly reachable. For localhost, open a cloudflared tunnel and pass the tunnel URL.
goal	string	The task each person tries, e.g. "complete the signup and reach the dashboard". Default: sign up and start using it.
count	integer	How many simulated people (default 5, max 10; without a key a free flow has up to 3, and defaults to 3). Each runs a full browser session and uses 5 credits.
access	object	Only for a site that is not open to everyone. vercel_bypass: a Vercel "Protection Bypass for Automation" secret for a protected preview. basic_auth: a username and password the site asks for in a browser prompt. headers: up to 4 headers sent only to this site. login: a test account (email or username, password, optional note) that the people sign in with to reach pages behind a sign-in. Used for this run only and never shown in results or reports. Ask the user before passing a password, and use a test account, not a real one.
devices	string	Which devices the agents use. "mix" (the default): about 4 in 10 on a phone-sized touch screen, the rest on a laptop. "phone" or "laptop" puts every agent on one.
audience	string	Who the simulated people should be, in plain words, e.g. "startup founders in SF" or "parents shopping for kids toys". If omitted, Mimiq picks the likely audience from the content.
max_steps	integer	Usually leave this out: each agent keeps going until it finishes the task or gets stuck, within the server's limit. Set it (10 to 200) only to cap a run on purpose.
review_id

{ "type": "object", "required": [ "url" ], "properties": { "url": { "type": "string", "format": "uri", "maxLength": 2000, "description": "Where the flow starts. Must be publicly reachable. For localhost, open a cloudflared tunnel and pass the tunnel URL." }, "goal": { "type": "string", "maxLength": 500, "description": "The task each person tries, e.g. \"complete the signup and reach the dashboard\". Default: sign up and start using it." }, "count": { "type": "integer", "default": 5, "maximum": 10, "minimum": 1, "description": "How many simulated people (default 5, max 10; without a key a free flow has up to 3, and defaults to 3). Each runs a full browser session and uses 5 credits." }, "access": { "type": "object", "properties": { "login": { "type": "object", "required": [ "email", "password" ], "properties": { "note": { "type": "string", "maxLength": 300 }, "email": { "type": "string", "maxLength": 320 }, "password": { "type": "string", "maxLength": 256 } }, "additionalProperties": false }, "headers": { "type": "object", "additionalProperties": { "type": "string", "maxLength": 2048 } }, "basic_auth": { "type": "object", "required": [ "username", "password" ], "properties": { "password": { "type": "string", "maxLength": 256 }, "username": { "type": "string", "maxLength": 256 } }, "additionalProperties": false }, "vercel_bypass": { "type": "string", "maxLength": 256 } }, "description": "Only for a site that is not open to everyone. vercel_bypass: a Vercel \"Protection Bypass for Automation\" secret for a protected preview. basic_auth: a username and password the site asks for in a browser prompt. headers: up to 4 headers sent only to this site. login: a test account (email or username, password, optional note) that the people sign in with to reach pages behind a sign-in. Used for this run only and never shown in results or reports. Ask the user before passing a password, and use a test account, not a real one.", "additionalProperties": false }, "devices": { "enum": [ "mix", "phone", "laptop" ], "type": "string", "description": "Which devices the agents use. \"mix\" (the default): about 4 in 10 on a phone-sized touch screen, the rest on a laptop. \"phone\" or \"laptop\" puts every agent on one." }, "audience": { "type": "string", "maxLength": 500, "description": "Who the simulated people should be, in plain words, e.g. \"startup founders in SF\" or \"parents shopping for kids toys\". If omitted, Mimiq picks the likely audience from the content." }, "max_steps": { "type": "integer", "maximum": 200, "minimum": 10, "description": "Usually leave this out: each agent keeps going until it finishes the task or gets stuck, within the server's limit. Set it (10 to 200) only to cap a run on purpose." }, "review_id": { "type": "string", "maxLength": 100, "description": "The review_id from a low-fit review or completed/running result. Bound to this URL and test kind." }, "timeout_seconds": { "type": "integer", "default": 1200, "maximum": 3600, "minimum": 60, "description": "How long to wait for the result. A run still going after this returns still_running with its report link." }, "confirm_audience": { "type": "boolean", "description": "Start with the reviewed people after showing them to the user." }, "audience_correction": { "type": "string", "maxLength": 1500, "description": "Not quite? Change who and rerun: use with review_id. Recruits a new crowd under the existing credit rules; starts unless the new crowd is wrong, otherwise returns a new review." }, "audience_fit_acknowledged": { "type": "boolean", "description": "Set only after telling the user about an audience-fit warning and they choose to continue." } } }