Integrations
Works with the tools your team already uses
AI Testing Probe connects to your tools through adapters, and its outputs stay there: test cases in Jira or Xray, tests in your Git repository, results as standard reports. Each row below shows the capability that uses the integration and that capability's current maturity label. Demos run Jira, Git and CI on simulators with the same adapter code.
By category
Integrations and their labels
Text badges, no logos. The label comes from the capability that uses the integration.
| Tool | What it does | Used by | Label |
|---|---|---|---|
| Requirements | |||
| Jira Cloud | Stories and acceptance criteria are read with a read-only account; fixes, questions and bugs are written by a project-limited bot account. | Requirement review | Demonstrated |
| Test management | |||
| Jira Test issues | Approved cases become Test issues with steps in the description, linked to the story. | Test case generation and Jira/Xray publishing | Demonstrated |
| Xray Cloud | Approved cases become manual Xray tests with structured steps. Optional. | Test case generation and Jira/Xray publishing | Demonstrated |
| Code and CI | |||
| GitHub | A GitHub App opens pull requests for generated tests and repairs. It never merges; people merge through your ruleset. | Template-bound Playwright code and pull request | Demonstrated |
| GitHub Actions | Runs the approved suite on pull request, push, deploy and nightly triggers, with no AI. | CI results into Control Tower and triage | Demonstrated |
| Azure DevOps | Boards, Repos, Pipelines and Test Plans. | Azure DevOps (Boards, Repos, Pipelines, Test Plans) | Roadmap |
| Browser and device | |||
| Playwright | Generated tests are plain Playwright TypeScript with your page objects. | Template-bound Playwright code and pull request | Demonstrated |
| Sauce Labs | Built to run the same tests on cloud browsers as a separate, non-blocking job. Adapter ready as a GitHub Actions job; not yet shown in a demo. | Sauce Labs cloud runs | Roadmap |
| FlaUI and Appium | Desktop and mobile code generation. | Desktop (FlaUI) and mobile (Appium) code generation | Roadmap |
| Model | |||
| Anthropic Claude | Swappable model gateway; Anthropic Claude today. Only one platform service holds the key. | Requirement review | Demonstrated |
| Azure OpenAI and other providers | New providers plug into the same gateway as adapters. | Azure OpenAI and other model providers | Roadmap |
| MCP clients | |||
| Claude Code and other MCP clients | Connect to the platform's MCP endpoint from the IDE with an API token. Tokens can submit drafts but cannot approve. | MCP client channel (e.g. Claude Code) | Demonstrated |
Not available today
- Other test management tools, such as TestRail or Zephyr. Test management is a port, but no other adapter exists today.
- Jira Data Center or Server. The Jira adapter targets the Jira Cloud REST API.
- Single sign-on (OIDC or SAML) for the portal. Users sign in with platform accounts and role-based access control.
- CI pipelines in GitLab, Jenkins or Azure DevOps. The approved suite runs in GitHub Actions today.
- Kubernetes or OpenShift deployment. The platform runs with Docker Compose on a single server today.
Each of these is Roadmap, with the date set by pilot demand. See them in the day-one matrix
Not on the list? Let's talk about adapter scope for your pilot.
Talk to usPrinciples
The same rules for every connection
-
Separate read and write identities
Reading and writing use different accounts or tokens. The write identity is limited to the project or repository the platform works in.
-
Credentials in one service
Model keys and write credentials go only to the core worker. The toolchain, the test runner and CI never receive them, and an automated check confirms it.
-
Secrets stay out of Git
Credentials live in the host's environment file or as CI secrets. Connection checks never print secret values.
-
Simulator or real, same code
Every external system sits behind a port. A local simulator can stand in for Jira, Git and CI during evaluation; switching is a configuration change.
Integrations setup: accounts and least-privilege permissions
Start with one story.
In a 30-minute call we look at your stack and a few of your stories, and tell you honestly whether a pilot fits.