Provides reviewed Kubernetes operations declared in YAML, enabling agents to perform backend checks and actions against Kubernetes environments.
Provides reviewed backend operations against PostgreSQL, declared in YAML, so agents can run typed database checks and operations as part of application testing.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Application testing with AI agents.
AgentQit exposes typed browser actions and reviewed backend operations through CLI and local MCP. The agent selects checks from requirements; the runtime validates inputs, enforces execution limits, and returns observations with evidence. UI, API, database, and messaging observations can share a task journal.
Availability: source preview on GitHub. The PyPI package is not published. Python 3.12+, macOS and Linux.
With uv installed:
uv tool install --python 3.12 'agentqit[browser] @ git+https://github.com/akimovWare/AgentQit.git@main'
qa browser install --headless-only
qa workspace init --directory ./my-qa --name sampleFor backend-only use, omit [browser] and the browser provisioning command.
Pin a commit instead of main for a repeatable source installation.
The workspace contains your operation YAML, profile, and credential references. Configure reviewed operations, QA endpoints, and an explicit credential provider using the workspace guide. Ordinary integrations require no Python code or runtime checkout; reusable catalogs and custom composition can use independent product plugins.
Configure your MCP host to launch:
agentqit-mcp --workspace /absolute/path/to/my-qaInteractive browser sessions use MCP. CLI commands provide backend execution, discovery, planning, diagnostics, and configuration. See installation and updates for package sources and upgrades.
Related MCP server: vouch
The synthetic example provides a local UI/API and a verification task:
git clone https://github.com/akimovWare/AgentQit.git
cd AgentQit
uv sync --frozen --extra browser
uv run qa browser install --headless-only
uv run python examples/orders/src/qa_product_orders/demo.pyOpen http://127.0.0.1:8765. Point your MCP host at the checkout's
.venv/bin/agentqit-mcp with --workspace /absolute/path/to/AgentQit/examples/orders.
Follow the Orders guide for mutation authorization,
expected behavior, and the deliberate frontend defect.
HTTP, PostgreSQL, Kafka, S3, and Kubernetes operations declared in YAML.
Bounded catalog search, typed input descriptions, and offline planning.
Interactive browser navigation, page observations, scoped element actions, and screenshots.
Isolated browser contexts, shared evidence journals, and private PNG/trace artifacts.
Explicit credential sources, recursive JSON redaction, TLS validation, and mutation admission.
Local workspaces and installable product plugins.
A completed operation is an execution fact. The agent compares observations with requirements to assess business behavior. Unknown write outcomes require state inspection before another attempt. Plugins execute trusted code; browser contexts provide task isolation rather than an OS sandbox. Read security boundaries.
Use | Guide |
Configure YAML and authentication | |
Install and update | |
Connect an application with an agent | |
Package an integration | |
Use browser sessions and evidence | |
Interpret results | |
Understand the components | |
Compare interface payloads | |
Contribute or release | |
Review development priorities |
One Python distribution, agentqit, provides qa, agentqit-mcp, and the runtime.
Browser support is optional; binaries are provisioned explicitly.
This server cannot be deployed
Autonomous cloud agent tasks: real browser + your tools, structured evidence-backed results.
Browser-based QA for AI-built software. Test pages with real browsers via agents.
Let your AI agent operate any website through a real, logged-in browser session.
Context and control for any AI. Agent identity, scoped access, and proof across tools.
