Reviews

Endpoints for triggering AI code reviews, querying review results, and accessing review details such as comments, files, rounds, and AI logs.

All review endpoints are under /api/reviews and require authentication.

---

List reviews

GET /api/reviews

Returns a paginated list of review history records, optionally filtered by project, status, or date range.

curl -X GET "https://your-domain.com/api/reviews?page=1&pageSize=20" \
  -H "Authorization: Bearer <api-key>"

Query parameters

page · number

Page number (1-indexed). Default: 1.

pageSize · number

Items per page. Default: 20.

projectId · string

Filter by project UUID.

status · string

Filter by review status. Values: pending, completed, failed.

startDate · string

ISO 8601 start date for date range filter.

endDate · string

ISO 8601 end date for date range filter.

Response 200

data.items · array

List of review history objects.

Show Review history fields

id · string

Review UUID.

projectId · string

Project UUID.

projectName · string

Human-readable project name.

mrIid · number

MR / PR number on the platform.

mrTitle · string

MR / PR title.

mrUrl · string

URL to the MR / PR on the platform.

status · string

pending | completed | failed

aiProvider · string

AI provider used for this review.

aiModel · string

Model identifier.

commentCount · number

Number of review comments posted.

triggeredBy · string

auto | manual

overallRating · string

Rating grade: A, B, C, D, or F.

summary · string

AI-generated review summary.

createdAt · string

ISO 8601 timestamp.

completedAt · string | null

ISO 8601 timestamp when the review finished.

data.stats · object

Aggregate statistics for the filtered result set.

meta.pagination · object

Standard pagination metadata.

---

Trigger a manual review

POST /api/reviews/trigger-manual-review

Manually queues an AI review for a specific merge request or pull request. Returns 202 Accepted with a task ID you can poll for status.

Requires the review:create permission.

curl -X POST https://your-domain.com/api/reviews/trigger-manual-review \
  -H "Authorization: Bearer <api-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "projectId": "<project-uuid>",
    "mrIid": 42,
    "sendToPlatform": true,
    "templateId": "<template-uuid>"
  }'

Request body

projectId · string · required

UUID of the project to review.

mrIid · number · required

Internal MR / PR number on the platform (the !42 or #42 number).

sendToPlatform · boolean

Whether to post review comments back to GitLab / GitHub. Defaults to the project setting.

templateId · string

UUID of the review template to use. Defaults to the project or global default.

Response 202

data.queued · boolean

true when the task was successfully queued.

data.task · object

Show Task fields

id · string

Task ID for polling.

status · string

queued | failed

sentToPlatform · boolean

Whether comments will be posted to the platform.

location · string

URL path to poll for task status.

---

Get review details

GET /api/reviews/:reviewId

Returns full details for a single review.

curl -X GET https://your-domain.com/api/reviews/<review-uuid> \
  -H "Authorization: Bearer <api-key>"

reviewId · string · required

UUID of the review.

Response 200

data.review · object

Full review record. Same fields as the list response items, plus baseCommitSha, headCommitSha, roundNumber, appliedRuleIds, and reviewedLines.

---

List review comments

GET /api/reviews/:reviewId/comments

Returns AI-generated comments for a review with optional filtering.

curl -X GET "https://your-domain.com/api/reviews/<review-uuid>/comments?severity=error" \
  -H "Authorization: Bearer <api-key>"

reviewId · string · required

UUID of the review.

filePath · string

Filter comments to a specific file path.

severity · string

Filter by severity. Values: info, warning, error.

limit · number

Maximum number of comments to return. Default: 20.

offset · number

Number of comments to skip for pagination. Default: 0.

Response 200

data.items · array

Show Comment fields

id · string

Comment UUID.

reviewId · string

Parent review UUID.

filePath · string

File path the comment references.

lineNumber · number

Line number in the diff.

severity · string

info | warning | error

message · string

The review comment text.

gitlabDiscussionId · string | null

GitLab discussion ID if the comment was posted.

createdAt · string

ISO 8601 timestamp.

---

List review files

GET /api/reviews/:reviewId/files

Returns the files that were reviewed.

reviewId · string · required

UUID of the review.

Response 200

data.files · array

Show File fields

id · string

File record UUID.

filePath · string

File path relative to repository root.

oldPath · string | null

Previous path for renamed files.

changeType · string

new | modified | deleted | renamed

summary · string | null

AI-generated summary for the file.

commentCount · number

Number of comments on this file.

---

Get file review progress

GET /api/reviews/:reviewId/files/progress

Returns per-file processing progress and prompt summary for an in-progress or completed review.

reviewId · string · required

UUID of the review.

Response 200

data.files · array

Array of file progress objects, each containing the file path and current processing status.

---

Retry a single file review

POST /api/reviews/:reviewId/files/:fileId/retry

Re-runs the AI review for a single file within an existing review. Requires review:create permission.

reviewId · string · required

UUID of the review.

fileId · string · required

UUID of the review file record to retry.

---

List review rounds

GET /api/reviews/:reviewId/rounds

Returns all review rounds. Multi-round reviews are used for differential re-reviews after new commits.

reviewId · string · required

UUID of the review.

Response 200

data.rounds · array

Show Round fields

id · string

Round UUID.

roundNumber · number

Sequential round number starting at 1.

isDifferential · boolean

Whether this round only reviewed changed files since the previous round.

createdAt · string

ISO 8601 timestamp.

---

List review feedbacks

GET /api/reviews/:reviewId/feedbacks

Returns user feedback records (positive / negative) for a review’s comments.

reviewId · string · required

UUID of the review.

Response 200

data.feedbacks · array

Show Feedback fields

id · string

Feedback UUID.

feedbackType · string

positive | negative

trackingStatus · string

pending | accepted | rejected | implemented

username · string | null

User who gave the feedback.

createdAt · string

ISO 8601 timestamp.

---

Get AI call logs

GET /api/reviews/:reviewId/ai-logs

Returns a log of every AI API call made during the review, including token usage and response times.

reviewId · string · required

UUID of the review.

limit · number

Maximum number of log entries to return. Default: 50, max: 200.

offset · number

Number of entries to skip. Default: 0.

provider · string

Filter by AI provider name.

status · string

Filter by call status (e.g., success, failed).

Response 200

data.items · array

Show AI log fields

id · string

Log entry UUID.

provider · string

AI provider name.

model · string

Model identifier.

status · string

Call outcome: success | failed.

durationMs · number

Round-trip duration in milliseconds.

promptTokens · number

Tokens in the prompt.

completionTokens · number

Tokens in the completion.

totalTokens · number

Total tokens consumed.

filePath · string | null

File being reviewed at call time.

createdAt · string

ISO 8601 timestamp.

---

Get execution logs

GET /api/reviews/:reviewId/logs

Returns runner / executor execution logs for a review.

reviewId · string · required

UUID of the review.

Response 200

data · object

Execution log data for the review’s runner task.

---

List all review tasks

GET /api/reviews/tasks

Returns a paginated list of all review queue tasks (both automatic and manual).

page · number

Page number. Default: 1.

pageSize · number

Items per page. Default: 20.

status · string

Filter by task status (e.g., pending, completed, failed).

---

Manual review helper endpoints

Get projects available for manual review

GET /api/reviews/manual/projects

Returns projects that are enabled and can have a manual review triggered.

Get manual review history

GET /api/reviews/manual/history

limit · number

Number of recent entries to return. Min: 1, max: 20, default: 5.

Get merge requests for a project

GET /api/reviews/manual/merge-requests

Fetches open MRs / PRs from the platform for use in the manual review dialog.

projectId · string · required

UUID of the project.

state · string

MR state filter: all | opened | merged | closed. Default: opened.

search · string

Search term to filter MRs by title.

Get recent manual tasks

GET /api/reviews/manual/tasks/recent

limit · number

Number of tasks to return. Default: 5.

Get manual task status

GET /api/reviews/manual/tasks/:taskId

Poll the status of a specific manual review task.

taskId · string · required

Task ID returned by the trigger endpoint.

Response 200

data.task · object

Task record including id, status, reviewId, and timestamps.

---

List review templates

GET /api/reviews/templates/all

Returns all available review prompt templates.

Response 200

data.templates · array

Show Template fields

id · string

Template UUID.

name · string

Template display name.

description · string | null

Optional description.

isDefault · boolean

Whether this is the global default template.

---

List review rules

GET /api/reviews/rules/all

Returns all configured review rules (file pattern matching, severity filters, etc.).

Response 200

data.rules · array

Show Rule fields

id · string

Rule UUID.

ruleName · string

Rule display name.

enabled · boolean

Whether the rule is active.

fileTypePatterns · array

Glob patterns for file extensions to match.

filePathPatterns · array

Glob patterns for file paths to match.

priority · number

Rule evaluation order (higher = earlier).