For agencies

Show clients the problem, then show them it is fixed.

Most clients do not argue with a video of someone failing to find the checkout. UX Surgery gives your team that evidence for every project.

Where it fits in your work

  • Before a redesign

    Record where visitors get stuck today, so the proposal starts from evidence instead of taste.

    ExampleSlide 2 of your proposal: “Checkout on mobile: 4 of 6 sessions stuck at shipping”, with the video clip.

  • During a retainer

    Run a monthly round on each client site and bring the ranked problems, the 3 worst first, to the meeting.

    ExampleFirst Monday of the month: one round per client, three problems on the agenda.

  • After a launch

    Re-test with the same profiles and tasks. The history shows what your work fixed.

    ExampleClient report: “Round 1: 4 of 6 blocked → Round 2: 0 of 6.”

  • Hand-off to developers

    Each problem comes with a fix prompt your developers or their coding agents can run directly.

    Examplefix-prompt.md attached to the ticket, with acceptance checks.

What the Agency plan adds

Before/after mockups of every problem
Every problem that allows it comes with a mockup of the smallest fix, so the client sees the change before anyone builds it.
Unlimited projects
One project per client or per product, with its own history. 500 credits a month cover, for example, 13 client sites at 36 sessions each (468 credits).
5 team members
Strategists, designers and developers in the same workspace.
Web and mobile apps (Android and iPhone)
Test your clients’ web apps and back offices behind a login, and their mobile apps on a phone, with test accounts they provide: Android apps, and iPhone apps on real devices (beta, 10 credits per session).
API, webhooks and CI
Launch studies from your deploy pipeline, get the results as JSON and a signed webhook when they are ready.

What you show the client (example)

Sessions blocked on the task “find the price”

  1. Round 14 of 6 Baseline
  2. Round 21 of 6 After adding “Pricing” to the menu
  3. Round 30 of 6 Fixed · 1 new minor issue

Agency

$299 per month

Many clients, one workspace.

  • 500 credits a month
  • Unlimited projects
  • Complete fix prompt
  • Before/after mockups of every problem
  • 5 team members
  • Web apps with login and Android apps
  • iPhone apps on real devices (beta)
  • API & webhooks for CI and client dashboards (beta)
  • White-label reports: coming

Example13 client sites × 36 sessions = 468 of 500 credits.

Choose Agency: Agency

API for your pipeline

Test every release before your client does.

Each deploy to a staging or preview URL can start a study on its own. Your pipeline gets the ranked problems, which ones are new since the last release, and the fix prompt for your coding agent.

  1. 1. Your pipeline launches a studyOne API call with the preview URL of the release, the commit and the branch. Retries never start a second study.
  2. 2. AI agents test the releaseThe same user profiles and tasks as last time, each study on its own machine, several at once.
  3. 3. You get a signed webhookNew problems, problems still open, fixed ones and any that came back, plus the video evidence and the fix prompt.

Works with GitHub Actions and GitLab CI: plain HTTP from any pipeline today, and a GitHub Action (beta, early access on request) that comments the top problems on each pull request.

Included in Agency. By default results are reported, not used to block a deploy: AI agents are not deterministic. You can choose to fail the build only when a serious problem is new or a fixed one reappears.

Read the API docs →

Launch a study from CI (curl)
curl -X POST https://app.uxsurgery.com/v1/studies \
  -H "Authorization: Bearer $UXS_API_KEY" \
  -H "Idempotency-Key: $GITHUB_SHA-$GITHUB_RUN_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "project_id": "'"$UXS_PROJECT"'",
    "target_url": "https://pr-128.staging.client.com",
    "config": { "preset": "standard", "formats": ["desktop", "mobile"] },
    "metadata": { "commit": "'"$GITHUB_SHA"'", "branch": "feature/checkout" }
  }'
GitHub Action (beta, not published yet)
# .github/workflows/ux-study.yml
on: pull_request
permissions:
  contents: read
  pull-requests: write   # one PR comment, updated on each run
jobs:
  ux:
    runs-on: ubuntu-latest
    steps:
      # GitHub Action (beta): not on the Marketplace yet.
      # Agency customers get the uses: reference with early access.
      - uses: UXS_ACTION
        with:
          api-key: ${{ secrets.UXS_API_KEY }}
          project: ${{ vars.UXS_PROJECT }}
          url: https://pr-${{ github.event.number }}.staging.client.com
          formats: desktop,mobile
          fail-on: none      # or new-critical / reappeared

Public pages, or with permission

You can study the public pages of any site, including competitors and prospects: the agents only read, like an anonymous visitor. Areas behind a login, web apps and mobile apps need the owner’s authorization, and web app tests need a verified domain. If you share results about a site you do not own, present them as AI-generated synthetic observations, in neutral terms.

Acceptable use policy

See where people get lost on your site.

Paste your URL. In a few minutes you get the 3 worst problems and one session video.

3 AI agent sessions on your site, the 3 worst problems and a video.