Now deployed at tinaa.hvs

Continuous Quality,
Intelligently Automated

TINAA MSP fuses Playwright automated testing with application performance monitoring through an AI-agent architecture — delivering real-time quality intelligence across every build and deployment.

90% Faster Test Creation
10x Coverage Increase
70% Cost Reduction
20+ MCP Tools

Everything quality requires

From test generation to performance monitoring, TINAA MSP closes the gap between code and confidence at every stage of delivery.

AI-Driven Test Generation

Describe a feature in natural language and TINAA generates executable Playwright test suites — with assertions, selectors, and edge cases included.

Learn more

APM Integration

Correlate test outcomes with real application performance data. Surface latency regressions, error spikes, and anomalies the moment they occur.

Learn more

MCP Protocol Native

Full FastMCP 2.8.0 support with 20+ documented tools. Works directly from VS Code, Cursor, and any MCP-compatible IDE through the agent protocol.

Learn more

Kubernetes Native

Helm charts, auto-scaling, and enterprise-grade deployment. Run at any scale with built-in observability for Prometheus and Grafana.

Learn more

Enterprise Security

OAuth2, SAML, RBAC, audit logging, Kubernetes secrets, and GitHub App authentication. Compliance-ready from day one.

Learn more

Multi-LLM Support

Connect OpenAI, Anthropic Claude, or local models. Intelligent prompt engineering adapts to your model's strengths automatically.

Learn more

Up and running in 60 seconds

Pull the pre-built Docker image and run your first AI-generated test in under a minute.

Pull the image

Download the latest TINAA MSP image from GitHub Container Registry.

Start the container

Launch with Docker Compose. Playwright, the MCP server, and the HTTP API start together.

Generate your first test

Hit the exploratory endpoint with any URL and receive a ready-to-run Playwright test.

Full User Guide
Terminal
# 1. Pull the latest image docker pull ghcr.io/aj-geddes/tinaa-playwright-msp:latest # 2. Start with Docker Compose curl -O https://raw.githubusercontent.com/aj-geddes/\ tinaa-playwright-msp/main/docker-compose.prod.yml docker-compose -f docker-compose.prod.yml up -d # 3. Verify it's healthy curl http://localhost:8765/health {"status":"healthy","version":"2.0.0"} # 4. Generate your first test curl -X POST http://localhost:8765/test/exploratory \ -H "Content-Type: application/json" \ -d '{"url":"https://github.com","focus":"nav"}'

Agent-based quality pipeline

TINAA MSP sits between your IDE and your infrastructure, translating intent into executable tests and correlating results with live APM data.

Deep-dive Architecture

Start shipping with confidence

Join teams who run continuous quality pipelines with TINAA MSP. Deploy in minutes, scale to thousands of tests.

Managed Service Platform  ·  v2.0

Continuous Quality,
Automated.

TINAA MSP is an autonomous quality platform that generates Playwright tests, monitors performance, computes a composite Quality Score, and gates deploys — all from a single AI-coordinated system.

TINAA MSP Dashboard — quality score, agent status, and recent test runs
1,400+ Tests generated
6 AI agents
14 MCP tools
0–100 Quality score
5 Web vitals tracked

Not just testing. Not just APM.
Both — fully autonomous.

Traditional testing tools require you to write and maintain tests manually. Traditional APM tools alert you after something breaks. TINAA MSP does both proactively: it reads your codebase, generates Playwright test playbooks automatically, monitors your live endpoints in real time, and computes a single Quality Score that tells you — and your CI pipeline — whether a deployment is safe.

Six specialised AI agents work under a central Orchestrator so that every product registered in TINAA is permanently under supervised quality watch.

Capability TINAA MSP Traditional tools
Auto-generate tests from code
Composite Quality Score
Synthetic monitoring + Web Vitals partial
Block deploys on quality failure
GitHub Check Runs on PRs
Claude Code / MCP integration

Everything quality requires, built in.

Six focused capabilities that work together so you ship with confidence.

Autonomous Testing

The Explorer and Test Designer agents analyse your repository structure and running application to auto-generate comprehensive Playwright playbooks. No manual test authoring required.

Quality Scoring

A composite 0–100 Quality Score weighs test health (40 %), performance (30 %), security (15 %), and accessibility (15 %) into one actionable number tracked over time.

APM & Web Vitals

Continuous synthetic monitoring tracks LCP, FCP, CLS, INP, and TTFB against configurable thresholds. The APM agent correlates performance regressions directly to deployments.

Deployment Gates

Define minimum Quality Score thresholds per environment. TINAA posts a GitHub Check Run on every PR — green to merge, red to block — keeping bad code out of production automatically.

GitHub Integration

Deep GitHub App integration: Check Runs on pull requests, deployment environment tracking, automatic issue creation on quality regressions, and webhook-driven test triggers on push.

Claude Code / MCP

A full MCP server exposes 14 tools so Claude Code can register products, trigger test runs, query Quality Scores, and manage alerts — all from your terminal without leaving your editor.



Three steps to autonomous quality.

From first registration to gated deploys in under an hour.

1
Register a product

Point TINAA at your repository and deployed environments. Provide a base URL and optional credentials. That is all the configuration required to start.

2
TINAA explores

The Explorer agent crawls your live application and analyses your codebase. The Test Designer generates a prioritised Playwright playbook. The APM agent begins endpoint monitoring.

3
Continuous quality

The Test Runner executes playbooks on every deploy. The Analyst computes the Quality Score. The Reporter gates your GitHub Check Run and sends alerts when thresholds are breached.


Agent-based, event-driven.

An Orchestrator coordinates six specialised agents over an async message bus.

Orchestrator
Coordinates all agents, manages state & scheduling
Explorer
Crawls app, maps routes
Test Designer
Generates Playwright playbooks
Test Runner
Executes tests, reports results
APM
Synthetic monitoring + Web Vitals
Analyst
Computes Quality Score
Reporter
Alerts, GitHub Checks, issues
  • API layer FastAPI, async Python 3.11+
  • Test engine Playwright (Chromium, Firefox, WebKit)
  • Primary store PostgreSQL + TimescaleDB (time-series metrics)
  • Cache / pub-sub Redis
  • Dashboard Web Components, Tailwind CSS, dark mode
  • MCP server FastMCP — 14 tools for Claude Code
  • Deployment Docker Compose or Kubernetes (Helm chart)
  • Alerts Slack, Teams, Email, PagerDuty, Webhooks, GitHub Issues

Up and running in minutes.

Docker Compose is the fastest path to a working TINAA MSP instance.

            bash
# 1 — Clone the repository
git clone https://github.com/aj-geddes/tinaa-playwright-msp.git
cd tinaa-playwright-msp

# 2 — Copy and edit environment variables
cp .env.example .env
#   Set GITHUB_TOKEN, DB credentials, etc.

# 3 — Start all services
docker compose up -d

# 4 — Confirm the API is ready
curl http://localhost:8000/health

# 5 — Open the dashboard
#   http://localhost:8000
  1. Clone & configure Copy .env.example to .env and fill in your database credentials and GitHub token.
  2. Start services docker compose up -d brings up the API, PostgreSQL, TimescaleDB, and Redis in one command.
  3. Register your first product POST to /api/v1/products with your repo URL and base application URL, or use the dashboard form.
  4. Connect GitHub Install the GitHub App and configure the webhook. TINAA will post Check Runs on your next pull request automatically.
  5. Connect Claude Code (optional) Add TINAA's MCP server to your Claude Code config for 14 AI-powered quality tools directly in your terminal.

Ready to automate quality?

Start with the getting-started guide or explore the full documentation to see how TINAA MSP fits into your delivery pipeline.