A project is one repository plus everything Coroid needs to work in it: build and test commands, context, policies and settings.
Create one from Projects → New, choosing a repository from a connected provider. If the repository you expect is missing, see Connect your code.
What happens on creation
Coroid clones the repository and analyses it to determine languages and frameworks, how it builds, how its tests run, and whether it can be containerised.
Check the result. See Build and test configuration — this is the configuration everything else depends on.
What a project owns
| Area | What it covers |
|---|---|
| Repository | Base branch, branch behaviour — see Repository and branch settings |
| Build and test | How the project compiles and verifies |
| Context | Documents, policies and references the agents read |
| Quality | Which policy and gates apply |
| Test targets | Environments and accounts for browser testing |
| Discoveries | What Coroid found in the codebase |
| Viability | File-size and pattern health signals |
One project per repository
Do not point two projects at the same repository to represent two workstreams. Use plans, sprints or lanes for that — see Tasks, plans, sprints and lanes — so both streams share one view of the code, one set of conventions and one context.
For a system spanning several repositories, create a project for each. Coroid can then coordinate work across them.
Scratch and preview projects
For experiments and throwaway work, a scratch project keeps noise out of your real ones. See Preview and scratch projects.
Deleting
Deleting a project removes its configuration, context and history from Coroid. It does not touch your repository — branches and pull requests Coroid created remain in your provider, under your control.
Deleted projects are recoverable for a period before removal becomes permanent. Retention is governed by your organization's settings — see Data retention.