> ## Documentation Index
> Fetch the complete documentation index at: https://doc.entelligence.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Set up Code Review
> Connect GitHub, GitLab, or Bitbucket, enable one repository, and verify the first Code Review in minutes.
Set up Code Review with one repository and one ready-for-review pull request. A small first run makes it easy to verify provider access, repository scope, review output, and automation before you expand.
## Before you start
You need:
* Workspace access to **Connectors** and **Code Review**
* Permission to install or authorize the GitHub, GitLab, or Bitbucket integration
* One repository visible to the workspace
* One non-draft test pull request
Supported providers include GitHub, GitLab, and Bitbucket, with supported self-hosted GitHub and GitLab installations.
## Connect source control
Open [**Connectors**](https://app.entelligence.ai/connectors) and select GitHub, GitLab, or Bitbucket.
Complete the provider flow. Start with one repository when repository-level scoping is available.
Open **Settings → Code Management** and confirm the repository appears. Reconnect or synchronize the provider if it is missing.
Use the dedicated guides for [Bitbucket Cloud](/bitbucket-cloud) and [self-managed GitLab](/gitlab-selfhosting).
## Enable the repository
Go to [**Code Review**](https://app.entelligence.ai/code-review) and select **Codebases → Repositories**.
Select **Add Repo** if it is not already listed.
Enable **PR Review**, then open **Show details**.
Start with **Medium priority and above**, inline comments, and a conservative comment limit. Keep automatic approval off during the pilot.
Confirm the repository and settings, then save the change.
## Run the first review
Open a non-draft pull request and mention the Entelligence bot with:
```text theme={null}
@entelligence review
```
If automatic review is enabled, the review can also start when the pull request opens, becomes ready for review, or receives a new commit.
## Confirm the setup works
A successful first run should:
* Appear under **Codebases → Active PRs**.
* Reference the correct repository and commit.
* Post the selected inline or summary output.
* Show only the enabled generated content.
* Recognize a reply in an Entelligence thread.
* Re-review a new commit only when **Auto Review on Every Commit** is enabled.
* Respect exclusions and the configured inline-comment limit.
## Tune the signal
After the first review is correct, tune one dimension at a time:
Adjust priority, comment limit, automatic triggers, and generated content.
Add guidelines, Learnings, repository Markdown, linked codebases, and exclusions.
Use a test pull request after each material change so you know which setting changed the output.
## First-run fixes
| Symptom | Check |
| -------------------------- | ------------------------------------------------------------------------- |
| Repository is missing | Provider authorization, repository scope, and **Code Management** sync |
| Pull request is missing | Correct repository, non-draft state, filters, and provider authentication |
| No review appears | **PR Review**, trigger behavior, bot mention, and provider permissions |
| Too many comments | Priority threshold and **Max Inline Comments** |
| Context is wrong | Assigned guidelines, Markdown context, Learnings, and linked codebases |
| Bot cannot post or approve | Provider write permissions and protected-branch policy |
## Next step
Learn [how to evaluate findings](/code-review/findings), or run a [local review before you push](/code-review/cli-review).