Overview

Deploy Verify is a post-deployment verification service that validates real user-facing behavior using AI-powered browser automation. After you deploy your web application, Deploy Verify launches a headless browser, follows the steps you define, and verifies that everything works correctly from a real user's perspective.

What makes Deploy Verify different?

• Natural language tests — Write instructions like "Click the login button" instead of document.querySelector('.btn-login').click()
• AI-powered browser — An AI agent understands your page and interacts with it like a real user
• Post-deploy verification — Automatically runs after every deployment via webhooks
• CI/CD integration — Trigger from GitHub Actions, GitLab CI, CircleCI, or any pipeline with a single curl command
• Screenshots & artifacts — See exactly what the browser saw during each test run

Quick Start Guide

Get up and running in four steps:

Key Concepts

Understanding the core data model helps you get the most out of Deploy Verify:

Creating a Site

A site represents a web application you want to verify after deployments. To create one:

1. Click "Sites" in the top navigation bar, or click "Add Your First Site" from the dashboard.
2. Click the "Add Site" button.
3. Fill in the required fields:
   • Site Name — A descriptive name for your site (e.g., "Production App", "Staging Environment").
   • Base URL — The root URL of your site (e.g., https://myapp.com). All test paths are relative to this URL.
4. Optionally add a description for your own reference.
5. Click "Create Site".

After creation, you'll be taken to the site detail page where you can set up your webhook and create tests.

Site Settings

Each site has the following configurable properties:

To edit a site, navigate to the site detail page and click the "Edit" button.

Authentication Config

If your site requires authentication (e.g., login credentials to access protected pages), you can configure an authentication config on the site. This is stored encrypted and is used by the browser automation agent to log in before running tests.

The auth config is provided as a JSON object when editing a site. The browser agent uses these credentials to perform authentication before executing your tests.

Managing Sites

The Sites page shows all your registered sites in a searchable grid view. Each site card displays:

• Site name and base URL
• Active/inactive status indicator
• Number of tests configured
• Recent run results

Click on any site card to view its detail page, where you can manage tests, view the webhook URL, and see recent run history.

Searching Sites

Use the search bar at the top of the Sites page to filter sites by name. The search updates in real-time as you type.

Deleting a Site

To delete a site, go to the site's edit page and use the delete option. Deleting a site also removes all associated tests, webhooks, runs, and artifacts.

Creating Tests

Tests define what Deploy Verify should do when verifying your site. Each test describes a user flow using natural language instructions and assertions.

To create a test:

1. Navigate to a site's detail page.
2. Click the "Create Test" button.
3. Fill in the test form:
   • Test Name — A descriptive name (e.g., "Login flow", "Homepage loads correctly").
   • Description — Optional notes about the test for your own reference. (e.g., "Confirming login functionality with test user").
   • Starting Path — The URL path where the test begins (e.g., /login, /). This is appended to the site's base URL.
   • Instructions — Natural language steps for the AI browser agent to follow. The easiest way to write these is with the AI Assistant.
   • Assertions — What should be verified after the instructions are executed.
4. Click "Save Test".

AI Assistant

The fastest way to write test instructions is with the AI writing assistant. Describe what you want to test in plain English, and it will generate detailed step-by-step instructions for you. You can then review and refine them.

Using the AI Assistant

1. In the test editor, expand the "AI Writing Assistant" panel.
2. Describe what you want to test in plain English (e.g., "Test that a user can log in and see their profile").
3. Click "Generate Instructions".
4. The AI will generate detailed step-by-step instructions for your test.
5. Review and edit the generated instructions as needed.

Writing Instructions

Whether you started with the AI Assistant or are writing from scratch, your instructions should be clear, step-by-step actions in plain English that tell the AI browser agent exactly what to do on your site.

Best Practices

• Be direct and specific — Say exactly what to do: "Click the Sign In button" not "Find a way to log in."
• Include concrete values — Use actual text: "Type '[email protected]' in the email field" not "Enter an email."
• One action per line — Break complex flows into sequential steps for reliability.
• Reference visible text — Refer to elements by their visible label text rather than CSS classes or IDs.

Example: Login Flow

Click the "Sign In" link in the navigation bar
Type "[email protected]" in the Email field
Type "password123" in the Password field
Click the "Log In" button
Wait for the dashboard page to load

Example: E-commerce Checkout

Click on the first product in the product grid
Click the "Add to Cart" button
Click the cart icon in the header
Verify the cart shows 1 item
Click "Proceed to Checkout"
Fill in the shipping form with test data
Click "Place Order"

Example: Form Submission

Click the "Contact Us" link in the footer
Type "John Doe" in the Name field
Type "[email protected]" in the Email field
Type "This is a test message" in the Message field
Click the "Send Message" button

Writing Assertions

Assertions define what should be true after your instructions have been executed. They are the "verification" part of Deploy Verify — if any assertion fails, the test fails.

Write assertions as plain English statements describing what the AI should check:

Example Assertions

The page should display "Welcome back, John"
The URL should contain "/dashboard"
A navigation menu should be visible
The "Log Out" button should be present
No error messages should be visible on the page

Assertion Types

Deploy Verify supports a wide range of assertion types for granular verification:

Smart Assertions Instant

Smart assertions are evaluated directly via Playwright's browser DOM APIs — they return results instantly with no AI processing. When your assertion text matches one of the patterns below, Deploy Verify skips the AI evaluation entirely.

Pattern Reference

CSS selectors must be wrapped in quotes (single or double) to trigger smart matching.

AI-Evaluated Assertions (Fallback)

Any assertion that doesn't match a smart pattern above is automatically sent to the AI model for evaluation. This is useful for subjective or visual checks:

The page looks like a login form
A navigation menu is visible with at least 5 items
The overall layout appears correct

Best Practices

• Prefer smart patterns — use the table above whenever the check can be expressed as a URL, text, or DOM query.
• Quote CSS selectors — element "#my-id" exists triggers smart matching; element #my-id exists does not.
• Use specific selectors — IDs (#logout-btn) are more reliable than class-based selectors that may change.
• Reserve AI assertions for subjective checks — "the page looks correct" or "no errors are visible" are good uses of AI evaluation.
• Check execution logs — smart assertions show as assertion_programmatic in the run detail page, confirming they bypassed the AI.

Test Settings

Each test has the following configurable properties:

Running Tests Manually

You can run any test manually without triggering the webhook. This is useful for developing and debugging tests.

1. Navigate to the test editor by clicking "Edit" on a test from the site detail page.
2. Click the "Run Test Now" button.
3. You'll be taken to the test runner page where you can watch the run progress in real-time.
4. Once complete, review the results including pass/fail status, step-by-step details, and captured artifacts.

Manual runs are tagged differently from webhook-triggered runs in the dashboard, so you can easily distinguish between the two.

Webhook Setup

Each site automatically gets a webhook when created. The webhook provides a unique URL that you can call to trigger a verification run across all active tests for that site.

Finding Your Webhook URL

1. Navigate to your site's detail page.
2. The webhook URL is displayed prominently at the top of the page.
3. Click the copy button to copy the URL to your clipboard.

The webhook URL has the format:

https://deployverify.com/api/webhook/YOUR_UNIQUE_TOKEN

Regenerating the Token

If your webhook token is compromised, you can regenerate it from the site detail page. Click the "Regenerate" button next to the webhook URL. This will invalidate the old token immediately — make sure to update your CI/CD pipelines with the new URL.

Triggering Webhooks

Trigger a verification run by sending a POST request to your webhook URL:

curl -X POST https://deployverify.com/api/webhook/YOUR_TOKEN

You can optionally include metadata about the deployment in the request body:

curl -X POST https://deployverify.com/api/webhook/YOUR_TOKEN \
  -H "Content-Type: application/json" \
  -d '{
    "commit": "abc123f",
    "branch": "main",
    "environment": "production",
    "actor": "jane.doe",
    "repository": "my-org/my-app"
  }'

Response

A successful trigger returns a 202 Accepted response with the run ID:

{
  "message": "Verification run queued successfully",
  "run_id": 42,
  "status_url": "https://deployverify.com/api/runs/42"
}

Use the status_url to poll for run results. Authenticate with the same webhook token using a Bearer header (see Run Status API).

GitHub Integration

Connect your GitHub repositories to enrich deployment data with commit details. When a verification run is triggered, Deploy Verify automatically fetches commit messages, authors, and links — giving you full context when diagnosing failures.

One-Click Connection

The recommended way to connect GitHub is through the one-click GitHub App flow:

1. Go to Site Settings and click Connect to GitHub
2. You'll be redirected to GitHub to authorize the Deploy Verify app
3. Select which repositories to grant access to
4. You'll be redirected back and can choose which repository to use for this site

The GitHub App uses short-lived installation tokens (1 hour) instead of long-lived personal access tokens, improving security. Multiple sites can share the same GitHub App installation with different repositories.

CI/CD Integration

Add Deploy Verify to your deployment pipeline to automatically verify your site after every deploy. Here are examples for popular CI/CD platforms:

GitHub Actions

- name: Trigger Deploy Verify
  if: success()
  run: |
    curl -X POST ${{ secrets.DEPLOY_VERIFY_WEBHOOK_URL }} \
      -H "Content-Type: application/json" \
      -d '{
        "commit": "${{ github.sha }}",
        "branch": "${{ github.ref_name }}",
        "actor": "${{ github.actor }}",
        "repository": "${{ github.repository }}"
      }'

GitLab CI

verify_deployment:
  stage: post-deploy
  script:
    - |
      curl -X POST $DEPLOY_VERIFY_WEBHOOK_URL \
        -H "Content-Type: application/json" \
        -d "{
          \"commit\": \"$CI_COMMIT_SHA\",
          \"branch\": \"$CI_COMMIT_REF_NAME\",
          \"actor\": \"$GITLAB_USER_LOGIN\",
          \"repository\": \"$CI_PROJECT_PATH\"
        }"
  only:
    - main

CircleCI

- run:
    name: Trigger Deploy Verify
    command: |
      curl -X POST $DEPLOY_VERIFY_WEBHOOK_URL \
        -H "Content-Type: application/json" \
        -d "{
          \"commit\": \"$CIRCLE_SHA1\",
          \"branch\": \"$CIRCLE_BRANCH\",
          \"actor\": \"$CIRCLE_USERNAME\",
          \"repository\": \"$CIRCLE_PROJECT_REPONAME\"
        }"

Laravel Forge

Add the following to the end of your Forge deployment script:

# Trigger Deploy Verify after deployment
curl -X POST https://deployverify.com/api/webhook/YOUR_TOKEN \
  -H "Content-Type: application/json" \
  -d "{\"environment\": \"production\", \"branch\": \"main\"}"

Generic / Other Platforms

For any platform, add a curl command after your deployment step. Store the webhook URL as an environment variable or secret:

#!/bin/bash
# After deployment completes...
curl -X POST "$DEPLOY_VERIFY_WEBHOOK_URL" \
  -H "Content-Type: application/json" \
  -d '{"environment": "production"}'

Webhook Metadata

Deploy Verify automatically extracts CI/CD metadata from your webhook requests. This metadata is stored with each run and displayed on the run detail page, in run lists, and in email notifications — helping you trace verification results back to specific deployments.

Supported Request Body Fields

Auto-Detected Headers

Deploy Verify also reads platform-specific headers for additional context:

• X-GitHub-Event and X-GitHub-Delivery — GitHub webhook headers
• X-GitLab-Event and X-GitLab-Token — GitLab webhook headers
• X-CircleCI-Signature — CircleCI webhook signature

Additionally, every run automatically records the triggered_at timestamp, request ip_address, and user_agent.

Understanding Runs

A run is a single verification execution that tests all active tests for a site. Runs can be triggered in two ways:

• Webhook — Automatically triggered by your CI/CD pipeline after a deployment.
• Manual — Triggered by clicking "Run Test Now" from the test editor.

When a run is triggered:

1. A run record is created with Pending status.
2. The run is queued for processing.
3. Each active test is executed as a Test Run within the run.
4. The AI browser agent visits the starting URL, follows the instructions, and checks the assertions.
5. Screenshots and other artifacts are captured during execution.
6. The run is marked as Passed (all tests pass) or Failed (one or more tests fail).

Run Statuses

Runs and individual test runs progress through the following statuses:

Test Run Details

Each run contains one or more test runs — one for each active test. A test run provides detailed results about the individual test execution:

Artifacts

Artifacts are outputs captured during test execution that help you understand what happened and debug failures. They are attached to individual test runs.

Artifact Types

Artifact Retention

Artifacts are stored for a limited retention period (default: 30 days). After the retention period expires, artifacts are automatically purged. Screenshots included in email notifications use signed URLs that also expire after the retention period.

Notification Preferences

Deploy Verify can notify you when your verification runs complete via email and Slack. Configure your preferences to control when and how you receive notifications.

To set your notification preferences:

1. Click your avatar in the top-right corner of the navigation bar.
2. Select "Notifications" from the dropdown menu.
3. Choose your preferred notification level.
4. Click "Save Preferences".

Email Notifications

Notification emails include key details about your verification run:

• Site name and overall pass/fail status
• Number of tests that passed and failed
• Total run duration
• For failed runs: which tests failed and the error messages
• Screenshots captured during the run (with signed URLs)
• A direct link to view the full run details in Deploy Verify

Slack Notifications Plus

Get real-time verification results delivered directly to your Slack workspace. Slack notifications are available on the Plus and Pro plans.

Setting Up Slack Notifications

1. Go to "Notifications" from the avatar dropdown menu.
2. Switch to the "Slack" tab.
3. Enable the "Send Slack notifications" toggle.
4. Create a Slack Incoming Webhook:
   1. Go to api.slack.com/apps and click "Create New App" → "From scratch".
   2. Navigate to "Incoming Webhooks" in the sidebar and activate them.
   3. Click "Add New Webhook to Workspace" and select the channel where you want notifications.
   4. Copy the webhook URL (it starts with https://hooks.slack.com/services/).
5. Paste the webhook URL into the "Webhook URL" field.
6. Choose your notification preference: "All Runs" or "Failures Only".
7. Click "Save Slack Settings".
8. Optionally click "Send Test Message" to verify the integration is working.

What's Included in Slack Messages

Each Slack notification includes:

• Overall pass/fail status with the site name
• Run number and test summary (e.g., "3 tests passed" or "1 failed, 2 passed")
• How the run was triggered (GitHub push, GitLab, manual, etc.)
• Total run duration
• For failed runs: the names of failing tests and their error messages
• A "View Run Details" button linking directly to the run in Deploy Verify

Notification Preferences

You can choose when Slack notifications are sent:

• All Runs — Get notified after every verification run completes, whether it passes or fails.
• Failures Only — Only get notified when a run fails or is cancelled. Useful for reducing noise in busy channels.

API: Trigger Webhook

Trigger a verification run for a site.

Parameters

Optional Request Body (JSON)

Response

202 Accepted

{
  "message": "Verification run queued successfully",
  "run_id": 42,
  "status_url": "https://deployverify.com/api/runs/42"
}

Error Responses

API: Webhook Info

Retrieve metadata about a webhook.

Response

200 OK

{
  "id": 1,
  "name": "Default Webhook",
  "is_active": true,
  "site": "My Production Site",
  "trigger_count": 47,
  "last_triggered_at": "2025-01-15T10:30:00Z"
}

API: Run Status

Get the status and results of a verification run.

Authentication

Authenticate using your webhook token in the Authorization header:

curl https://deployverify.com/api/runs/42 \
  -H "Authorization: Bearer YOUR_WEBHOOK_TOKEN" \
  -H "Accept: application/json"

This is the same token from your webhook URL. Authenticated site owners can also access this endpoint via their browser session.

Response

200 OK

{
  "id": 42,
  "status": "passed",
  "site": "My Production Site",
  "started_at": "2025-01-15T10:30:05Z",
  "completed_at": "2025-01-15T10:31:22Z",
  "duration_ms": 77000,
  "passed_count": 3,
  "failed_count": 0,
  "error_message": null,
  "ai_summary": "All tests passed successfully.",
  "test_runs": [
    {
      "id": 101,
      "test": "Login Flow",
      "status": "passed",
      "duration_ms": 25000,
      "steps_passed": 5,
      "steps_failed": 0,
      "assertions_passed": 3,
      "assertions_failed": 0,
      "error_message": null,
      "failed_at_step": null
    }
  ]
}

Profile Settings

Access your profile settings by clicking your avatar in the top-right corner and selecting "Profile".

From the profile page you can:

• Update your name and email address
• Change your password
• Delete your account

Contact Us

Have a question, found a bug, or need help getting set up? Reach out to our support team and we'll get back to you as soon as possible.

Reach out via our contact page and we'll get back to you as soon as possible.