Initial version

Changes

• HTTP interceptor
• Console logs interceptor
• User interaction interceptor
• Navigation interceptor

Added screen recording, screenshots. More compact layouts, a lot of bugfixes.

Changes

• Implemented screen recording
• Implemented screenshots
• Improved network cards UI

Added Websocket support

Changes

• WebSocket interceptor

Adding screenshots and screen recordings to the logs. Attaching static, styled html in generated logs archive. Automatically adding all the screenshots and screen recordings taken to the generated logs archive.

Changes

• Automatically add screenshots to generated logs archive
• Automatically add screen recordings to generated logs archive
• Show screen recordings in logs
• Show screenshots in logs

Automatically generate end-to-end tests from user interaction events

Changes

• Automatically generate Cypress tests
• Automatically generate Playwright tests
• Automatically generate Puppeteer

Advanced HTTP filters, device info, better E2E generation, GraphQL

Changes

• Advanced HTTP filters - eliminate the noise from the logs and make it easier to find what matters
• End-to-end - More widely used test attributes support data-cy, data-test, data-testid in end-to-end test generator
• Device info - resolution, user agent will be exported with the logs to make debugging even easier
• GraphQL support - GraphQL requests will be parsed and formatted so it's easier to see queries, mutations and subscriptions
• Better, more responsive UI

Hiding sensitive data, advanced Websocket filters, Monito API

Changes

• Hiding sensitive data - now you can specify which parts of the logs you want to remove before extracting
• Advanced Websocket filters - filter out not needed Websocket frame types
• Monito API - if you use Monito on daily basis you can send custom events from your web application and they will automatically appear in the logs
• Bugfixes

Simplified design, refreshed website, Stripe Checkout integration

Changes

• Refreshed website, blog and Stripe Checkout integration instead of Gumroad
• Refreshed extension design, new fonts, better UX
• Desktop Capture extension API usage for handling screen/window/tab recording
• Bugfixes

Monito becomes an AI QA platform for plain-English, real-browser testing

Changes

• Run reusable test scenarios in a real browser and review pass/fail verdicts with screenshots, console errors, network evidence, and session replay
• Discover application flows and review draft scenarios before saving them
• Group scenarios into suites and run them against production, staging, or preview environments
• Automate testing with schedules, deployment webhooks, GitHub Actions, and short-lived GitHub OIDC authentication
• Manage projects, scenarios, runs, discovery, automations, and evidence through the agent-friendly Monito CLI

An MCP server, so AI coding agents can run tests and read bug reports

Changes

• Hosted MCP server at https://www.monito.dev/api/mcp for Claude Code, Cursor, VS Code, Claude, and ChatGPT, with OAuth sign-in in the browser and a Connected apps list in Settings to review and revoke access
• Local MCP server in the CLI with monito mcp (or npx -y @monitodev/cli mcp), using your CLI login and adding tools for setting up CI
• Run launches accept a requestId, so a retried launch returns the original run instead of charging twice (--request-id in the CLI)

One command to check a change, failure emails with causes, and the cost before you spend

Changes

• monito check <url> "<what should happen>" (and the MCP check tool) returns one verdict with a report link; monito check save keeps it as a test
• The MCP server's default tools are now the seven a coding agent needs; ?toolsets=all restores the full set
• In GitHub Actions, --github-check adds a Monito check to the commit and --github-comment keeps at most one PR comment, only when something fails
• One email when an automated suite fails, naming each failing test and why it failed; a Settings toggle turns run alerts off
• Launch controls show the cost and your balance first; finished runs say what was charged
• A Defects page per project, Connect Slack from Settings, and an export of a project's scenarios as plain-English Markdown

A first verdict in minutes, why each run failed, request headers for protected previews and a suite on every Vercel preview

Changes

• Onboarding runs a free first test before anything else and asks whether your app needs a login
• Every failed run says why — app bug, test needs updating, site unreachable, or Monito's error (not charged) — on the report, in run lists, the CLI and MCP
• Environments can send secret request headers, such as Vercel's protection bypass, only to their own HTTPS origin
• A signed Vercel deployment webhook runs your suite against every preview deployment
• Status pages are reachable on <slug>.monito.page, and the journeys editor lists your tests again
• monito ci knows when a project uses the Vercel webhook and won't silently switch it back

Check your localhost from your coding agent, runs that finish, and every reliability number linked to its real run

Changes

• monito check http://localhost:3000 … (or --local 3000) checks your dev server in a real browser through a per-run tunnel, from the CLI or the local MCP server; a tunnel that never connects is free
• Long tests run to their full time budget, and runs stopped by the budget or a limit on Monito's side are not charged
• Improve in AI assist never narrows what a test checks
• A hint when a long flow has a short timeout, in the editor, assist, the CLI and MCP
• /reliability links every result to the public report of the run behind it
• A new account's first test checks the page you entered, with the standard 5-minute budget
• "Bugs found" emails go only to scheduled, webhook and CI runs, at most five an hour
• monito check --max-duration sets the run's time budget