Changeset
Changesets are AI-powered code improvement workflows in Artemis. They help you analyze, fix, and validate code changes with automation and AI assistance.
What you can do with Changesets?
- Generate improvements with custom prompts or pre-defined templates
- Combine versions of each target into your own chosen set
- Validate automatically with testing and benchmarks
- Create branches and pull requests directly from the Artemis interface
- Track the full history of all code modifications
You can access the changeset tab from your Targets panel.
How to Create a Changeset?
Step 1: Open Targets
Navigate to:
Project → Optimise → Targets
Step 2: Start a New Changeset
Click Create changeset at the top of the panel.
Step 3: Choose Your Workflow
You have two options when creating a changeset:
If you choose Start New Task:
Create AI-generated code improvements using prompts or templates.
Examples:
- Fix a bug and add a validation test
- Optimise a top performance bottleneck
- Auto-generate repository documentation
- Harden a selected security issue
If you choose Select Targets:
Pick specific code targets and select which version you want to use for each.
Here you can:
- Mix and match chosen versions for multiple targets
- Validate the changes automatically
- Create a branch and open a pull request