GitLab

Use the GitLab Connector to ingest repository files and, optionally, issues from a GitLab project into the mAItion knowledge base.

What It Does

fetches files from a GitLab project’s repository at a given branch, tag, or commit
optionally limits ingestion to a sub-directory or a single file
optionally ingests the project’s issues, with filtering by state, labels, assignee, author, milestone, and more
runs ingestion on configurable schedules

Authentication

Authenticate with a GitLab Personal Access Token scoped to read_api. Supports both GitLab.com and self-hosted GitLab instances via gitlab_url.

Known Limitation

The underlying reader re-downloads all repository file content from the GitLab API on every scheduled run. The vector store’s checksum-based deduplication still prevents re-embedding unchanged files, but it does not reduce GitLab API calls for the files connector.

Required Environment Variables

Set these in .env.rag:

GITLAB1_URL: base URL of your GitLab instance (example: https://gitlab.com)
GITLAB1_TOKEN: Personal Access Token with read_api scope
GITLAB1_SCHEDULES: ingestion interval in seconds (default: 3600)

`config.yaml` Example

sources:
  - type: "gitlab"
    name: "gitlab1"
    enabled: true  # optional, default: true
    config:
      gitlab_url: "${GITLAB1_URL}"
      personal_token: "${GITLAB1_TOKEN}"
      project_id: 12345678              # integer project ID, required
      ref: "main"                       # optional, branch/tag/commit, default "main"
      #path: "docs"                     # optional, limit to sub-directory
      #file_path: "README.md"           # optional, single file only
      recursive: true                   # optional, default true
      files_iterator: true              # optional, use iterator pagination to fetch all files (default true); set to false to limit to 20 files (GitLab API default page size)
      include_issues: false             # optional, default false
      #issues_state: "opened"           # optional: opened/closed/all, default "opened"
      #issues_labels: "bug,docs"        # optional, comma-separated
      #issues_assignee: "username"      # optional
      #issues_author: "username"        # optional
      #issues_milestone: "v1.0"         # optional
      #issues_search: "keyword"         # optional
      #issues_get_all: true             # optional, fetch all pages, default true; set to false to limit to 20 issues (GitLab API default page size)
      #issues_scope: "created_by_me"    # optional: created_by_me/assigned_to_me/all
      #issues_type: "issue"             # optional: issue/incident/test_case/task
      #issues_confidential: false       # optional
      #issues_non_archived: true        # optional
      #issues_iids: [1, 2, 3]           # optional, filter by specific issue IDs
      #issues_created_after: "2024-01-01T00:00:00Z"   # optional, ISO-8601
      #issues_created_before: "2024-12-31T23:59:59Z"  # optional, ISO-8601
      #issues_updated_after: "2024-01-01T00:00:00Z"   # optional, ISO-8601
      #issues_updated_before: "2024-12-31T23:59:59Z"  # optional, ISO-8601
      schedules: "${GITLAB1_SCHEDULES}"

`.env.rag` Example

GITLAB1_URL=https://gitlab.com
GITLAB1_TOKEN=glpat-xxxxxxxxxxxxxxxxxxxx
GITLAB1_SCHEDULES=3600

Configuration Reference

Field	Required	Default	Description
`enabled`	no	`true`	Set to `false` to skip this source entirely
`gitlab_url`	yes	—	Base URL of the GitLab instance
`personal_token`	yes	—	Personal Access Token with `read_api` scope
`project_id`	yes	—	Integer project ID to ingest files (and optionally issues) from
`ref`	no	`main`	Branch, tag, or commit to read files from
`path`	no	—	Limit file ingestion to a sub-directory
`file_path`	no	—	Limit file ingestion to a single file
`recursive`	no	`true`	Recurse into sub-directories when listing repository files
`files_iterator`	no	`true`	Use iterator pagination to fetch all files; set `false` to limit to 20 files (GitLab API default page size)
`include_issues`	no	`false`	Set to `true` to also ingest issues
`issues_state`	no	`opened`	`opened`, `closed`, or `all`
`issues_labels`	no	—	Comma-separated label filter
`issues_assignee`	no	—	Filter by assignee username
`issues_author`	no	—	Filter by author username
`issues_milestone`	no	—	Filter by milestone title
`issues_search`	no	—	Free-text search filter
`issues_get_all`	no	`true`	Fetch all pages of results; set `false` to limit to 20 issues (GitLab API default page size)
`issues_scope`	no	—	`created_by_me`, `assigned_to_me`, or `all`
`issues_type`	no	—	`issue`, `incident`, `test_case`, or `task`
`issues_confidential`	no	—	Filter by confidential flag
`issues_non_archived`	no	`true`	Exclude issues from archived projects
`issues_iids`	no	—	List of specific issue IIDs to fetch
`issues_created_after` / `issues_created_before`	no	—	ISO-8601 date bounds on issue creation
`issues_updated_after` / `issues_updated_before`	no	—	ISO-8601 date bounds on issue updates
`schedules`	no	`3600`	Ingestion interval in seconds

Multiple GitLab Sources

Add more sources entries (gitlab2, gitlab3, etc) with separate env vars per source.

Getting started

Features

Connectors

Tools

Configuration

What It Does

Authentication

Known Limitation

Required Environment Variables

`config.yaml` Example

`.env.rag` Example

Configuration Reference

Multiple GitLab Sources

​What It Does

​Authentication

​Known Limitation

​Required Environment Variables

​config.yaml Example

​.env.rag Example

​Configuration Reference

​Multiple GitLab Sources

What It Does

Authentication

Known Limitation

Required Environment Variables

`config.yaml` Example

`.env.rag` Example

Configuration Reference

Multiple GitLab Sources