Dokkimi

Dokkimi is a testing framework designed for integration, end-to-end (E2E), and visual regression testing, enabling users to test applications without relying on staging environments. It provides production-like, isolated environments on demand, allowing tests to be run with dedicated services, databases, and browsers for each test run, which eliminates shared state and prevents test corruption.

The platform supports driving real browsers for UI testing, enabling actions such as clicking, typing, navigating, and making assertions based on what the user sees. Visual regression testing is facilitated through screenshotting steps and comparing them against baselines to detect visual changes before deployment. Dokkimi also captures every HTTP call between services, allowing users to assert on the exact data sent and received. It can intercept calls to external APIs, such as Stripe, Auth0, and Twilio, and return controlled responses, removing the need for test accounts. Database seeding and querying are supported for Postgres, MySQL, MongoDB, or Redis, with the ability to intercept and log every query for debugging and to make direct assertions on database content.

Dokkimi is built for integration with AI agents, shipping with a full MCP server that allows AI tools to validate, run, inspect, and debug tests as native tool calls. On installation, it auto-registers with Claude Code, Cursor, and GitHub Copilot, enabling agents to write and manage tests immediately. Test definitions and configurations are managed using structured YAML or JSON files, which AI agents can read, write, and iterate on. The tool provides complete structured output to help AI diagnose failures, analyze traffic, and improve tests.

Installation is available via Homebrew and npm, and Dokkimi runs locally or in continuous integration (CI) environments using Docker. It offers a free tier for local and CI use, with unlimited test definitions and runs, E2E UI testing, visual regression, traffic interception, mocks, database assertions, GitHub Action integration, and AI agent compatibility. 0. A cloud offering with additional features such as parallel execution, results dashboards, flaky test detection, and regression tracking is planned for the future.