Tutorials

Import your first report

Import your first report starts TestLookup on your own machine. It then uploads a small JUnit file with two tests to a test project. JUnit is a common test report format.

You follow the report from the moment TestLookup accepts it, through reading the file, to a finished test run. The tutorial shows why you keep both IDs that TestLookup sends back. It also shows that “accepted” means “queued”, not “done”.

The account that sends the report needs the QA engineer role or higher. TestLookup refuses a viewer.

Details: TestLookup confirms an accepted report with HTTP 202, and refuses a viewer with HTTP 403.

Connect a CI job

Continuous integration (CI) is the service that builds and tests your code on every change, such as GitHub Actions or Jenkins. Connect a CI job takes you through these steps:

1. Install the TestLookup command-line tool.
2. Save the server address, the key and the project ID as secrets in your CI system.
3. Upload a report with a build name that stays the same if the job retries.
4. Wait until TestLookup has saved the results.
5. Pass or fail the job based on the run’s real failures.

The tutorial also shows a separate command that passes or fails a job based on a release’s saved decision. The key’s owner needs the QA engineer role or higher to send results.

Details: the run check is testlookup ci-verdict, and the release check is testlookup release gate.

Review an AI report

Review an AI report finds a report under Inbox › Approvals. You check its evidence and how it was made, then accept or reject it. The tutorial then shows how the review gate affects exports.

Four things stay separate: a suggested classification, an accepted report, an issue in another tool and a release decision.

Configure local AI

Configure local AI starts the optional AI model services on your own machine. It checks which models are installed and which ones TestLookup will use. Then it confirms that an analysis works.

Stopping TestLookup keeps your data and downloaded models. A separate clean-up command deletes them.

make dev-llm              # start TestLookup with the local AI services
make stop                 # stop, and keep your data and models
make clean CONFIRM=yes    # delete all data and downloaded models

Details: by default, TestLookup calls the model qwen2.5:7b-instruct.

Requirements and expected results

Each tutorial names the tools, account and setup it needs, the commands to run and the result you should see. It also says how to recover and how to clean up.

• Use a test project and a TestLookup installation that you can throw away.
• Write down where a step failed.
• Stop TestLookup rather than running commands that delete data.

A tutorial result shows that the steps work. It doesn’t measure capacity, test a deployment or show how accurate a model is.

Each tutorial ends with a run recorded on September 21, 2026, with an earlier development version of TestLookup. Details may differ on the current version.

For more, read the workflow guide and the development guide on GitHub.

Start with Import your first report.