Welcome to the
Agentic QA Era
aqua is a QA planning and execution service for AI coding agents. Describe what to test, and AI handles the rest — from generating a plan to running it and sharing results.
Dashboard Login Flow
completedNormal Login
Login Failure
Login Page UI
Works with MCP-compatible AI agents
Built for teams shipping with AI
AI makes coding fast — but QA is still manual
Code ships in minutes with AI agents, but verifying it works still takes hours of manual clicking
Each pull request changes different things — figuring out what to test every time is exhausting
Point the AI at a pull request and it reads the diff to generate a QA plan tailored to the changes. API tests, browser checks, edge cases — all covered automatically.
Tried AI for QA — but it didn't quite work
You asked an AI agent to test your app, but the results vary every time — coverage is inconsistent and instructions aren't reusable
Letting the agent drive the browser directly floods its context with page content and screenshots — leaving less room for reasoning. Flaky selectors and credential exposure add to the headache.
aqua separates planning from execution. Plans are versioned and reusable. Browser automation runs on a dedicated engine. Credentials stay in secret providers — never exposed to the AI.
How It Works
Describe what to test, then let the AI handle the rest
Describe What to Test
Tell the AI agent what to test in natural language.
- Point at a pull request for auto-generated tests
- Describe API endpoints, user flows, or error cases
- The agent reads your codebase to understand the app
AI Creates QA Plan
The agent builds a structured plan with scenarios, steps, and assertions.
- Analyzes your codebase and PR diff to build test scenarios
- Review the plan and give feedback to refine it
- Covers both HTTP API tests and browser interactions
Execute & Collect Results
Run the plan via your coding agent or CLI.
- HTTP logs, screenshots, and DOM snapshots captured automatically
- Assertion results reported for every step
- View detailed results on the web dashboard with your team
Learn & Iterate
The AI saves insights to project memory for the next session.
- Reliable selectors, auth flows, and tips accumulate over time
- Plans are reusable — re-execute on every deploy
- Team members share knowledge across sessions
Reusable Plans
Write a plan once, re-run it every time you push a fix to a PR. No need to regenerate — just iterate on your code and re-execute.
Environment Management
Switch between local, staging, and production with a single flag. URLs, variables, and secrets are managed per environment.
Project Memory
The AI remembers what it learns — reliable selectors, auth flows, API patterns — so each QA session builds on the last. No more re-explaining your app every time.
MCP-Native
Built for the Model Context Protocol. Works with Claude Code, Gemini CLI, OpenCode, Cursor, and any MCP-compatible AI agent.
Security & Privacy
Your credentials and test data are protected by design
Secrets Stay Local
Credentials resolve locally on your machine through providers like 1Password, AWS Secrets Manager, GCP Secret Manager, and HashiCorp Vault. They are never sent to AI agents or the aqua service.
Automatic Masking
Before test artifacts reach the server, sensitive data — authorization headers, cookies, password fields, and known secret values — is automatically detected and masked.
Access Control & Audit
Organization-level roles (Owner, Admin, Member, Viewer) control who can do what. All operations are recorded in audit logs for full traceability.
Need complete data sovereignty? Self-host aqua on your own infrastructure — your data never leaves your environment.
FAQ
Frequently asked questions about aqua
Personal use is free. Commercial use will be a paid plan in the future. During the current beta period, all features are available at no cost.
With aqua, QA plans and execution are separated. Plans are reproducible and reusable across runs, and credentials or secrets never leave your local machine — they don't need to be passed to the AI agent.
aqua works with any AI agent that supports the Model Context Protocol (MCP) — including Claude Code, Gemini CLI, OpenCode, Cursor, Windsurf, and Antigravity. You can also use the CLI directly without an AI agent.
Yes. The aqua-cli execute command can run QA plans from the command line, making it easy to integrate into GitHub Actions, GitLab CI, or any CI/CD pipeline.
Absolutely. aqua is not a replacement for unit tests or integration tests. It focuses on end-to-end QA scenarios — API flows, browser interactions, and cross-service verification — complementing your existing test infrastructure.
Ready for Agentic QA?
Let AI agents plan and execute your QA — built for the agentic era.