Where does GitHub look for PULL_REQUEST_TEMPLATE.md?

GitHub checks the .github directory, the repository root, and the docs directory on the default branch. The .github directory is the conventional choice. You can also keep multiple templates in .github/PULL_REQUEST_TEMPLATE/ and select one via the template query parameter.

Can I have different templates for different kinds of changes?

Yes. GitHub supports a PULL_REQUEST_TEMPLATE directory with multiple Markdown files, chosen with the template query parameter in the compare URL. Generate one template per change type here — for example a compact one for fixes and a release-depth one for migrations.

Do pull request templates help AI code review too?

Yes. An AI reviewer compares the implementation against the stated intent, so a description that states intent clearly gives it something concrete to verify. The same structure that helps a human reviewer helps the model.

All tools

Pull Request Template Generator

Generate a compact GitHub PULL_REQUEST_TEMPLATE.md that asks authors for the context reviewers actually need.

Configure the output

The result updates immediately and stays in this browser.

PULL_REQUEST_TEMPLATE.md

## Summary

<!-- What changed, and why is this the right change? -->

## Review focus

<!-- Where should reviewers spend their attention? -->

## Related issue

Closes #

## Test plan

- [ ] Automated tests
- [ ] Manual verification
- [ ] Regression or failure-path check

## Screenshots or recordings

<!-- Include before/after evidence for visible changes, or remove this section. -->

## Rollout and rollback

<!-- Note migrations, flags, deployment order, monitoring, and rollback steps. -->

## Security and data impact

<!-- Note trust-boundary, permission, secret, or sensitive-data changes. -->

## Author checklist

- [ ] The PR is focused and its description matches the implementation.
- [ ] Tests and documentation are updated where needed.
- [ ] Breaking changes, migrations, and operational risks are explicit.
- [ ] No credentials, private data, or unrelated generated files are included.

Where should the pull request template live?

Commit the generated file as PULL_REQUEST_TEMPLATE.md in the repository root, docs directory, or .github directory. GitHub will place its contents into new pull request descriptions after the template reaches the default branch.

The generator offers three depths because template size should follow repository risk. A compact template suits fast-moving product repos where a sentence of intent and a test note are enough. The standard depth adds rollout and screenshot prompts. The release depth adds the operational sections — deployment order, rollback, monitoring — that regulated or high-traffic services need on every change.

The most common template failure is not absence but bloat: ten mandatory sections that authors fill with 'N/A' until reviewers stop reading descriptions entirely. Every section in the generated output earns its place by changing a review decision. If a section would not change how someone reviews the change, delete it.

Keep the template short enough that authors complete it honestly. Remove sections that do not change review decisions, and use repository-specific examples where ambiguity is common.

Frequently asked questions

Where does GitHub look for PULL_REQUEST_TEMPLATE.md?: GitHub checks the .github directory, the repository root, and the docs directory on the default branch. The .github directory is the conventional choice. You can also keep multiple templates in .github/PULL_REQUEST_TEMPLATE/ and select one via the template query parameter.
Can I have different templates for different kinds of changes?: Yes. GitHub supports a PULL_REQUEST_TEMPLATE directory with multiple Markdown files, chosen with the template query parameter in the compare URL. Generate one template per change type here — for example a compact one for fixes and a release-depth one for migrations.
Do pull request templates help AI code review too?: Yes. An AI reviewer compares the implementation against the stated intent, so a description that states intent clearly gives it something concrete to verify. The same structure that helps a human reviewer helps the model.

Try Scopy AI on your next pull request

Accurate, open-source AI code reviewer that understands your project. Self-host it or start in the cloud.

Star on GitHub Get started