Skip to main content
Glama

Model Registry

gitlab_model_registry
Read-onlyIdempotent

Download ML model files from GitLab Model Registry for inference, evaluation, or vendoring into build pipelines.

Instructions

Example: {"action":"download","params":{...}} For the params schema of any action, read the MCP resource gitlab://schema/meta/gitlab_model_registry/.

Download ML model package files from the GitLab Model Registry. Read-only — cannot publish or delete model versions through this tool. The underlying GitLab API requires a Premium/Ultimate plan on the target instance (server enforces it with 403); the tool itself is always registered and is not gated by GITLAB_ENTERPRISE. When to use: pull a model artifact (.pkl, .onnx, .safetensors, .bin, .gguf, etc.) attached to a registered model version, e.g. for inference, evaluation or vendoring into a build pipeline. NOT for: generic packages (use gitlab_package), container images (use gitlab_package registry_), release attachments (use gitlab_release link_), training jobs or experiment tracking, model publishing or versioning (not yet exposed through MCP).

Returns:

  • download: {file_name, model_version_id, size, content_base64} — binary content is base64-encoded; large models can produce very large responses. Errors: 404 (hint: project_id, model_version_id and path are model-registry-scoped; verify in the GitLab UI under Deploy → Model registry), 403 (hint: requires Reporter+ on the project and a Premium/Ultimate plan), 400 (hint: filename must match an asset attached to the version).

  • download: project_id*, model_version_id*, path*, filename*. Returns base64-encoded file content.

    • project_id (string | int, required) — numeric ID or URL-encoded full path of the project that owns the registered model.

    • model_version_id (int, required) — registered model version ID; visible in the GitLab UI under Deploy → Model registry → → Versions.

    • path (string, required) — package-relative directory of the asset (use '/' for the package root, otherwise e.g. 'artifacts/' or 'weights/').

    • filename (string, required) — exact asset filename within the package, including extension (e.g. 'model.safetensors', 'config.json').

    • Any unrecognized parameter name is rejected by the meta-tool router (strict unknown-field validation); only the reserved meta key 'confirm' is stripped before unmarshalling.

See also: gitlab_package (generic / npm / maven / conan / pypi / nuget / container packages), gitlab_release (asset links per release), gitlab_repository (raw files in the repo).

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
actionYesAction to perform. Pick exactly one of the values in `enum`. Each action expects its own `params` object — see the tool description for the per-action parameter list.
paramsNoAction-specific parameters as a JSON object. Required and optional fields differ per action; consult this tool's description for the chosen action. Send only the fields documented for that action — unknown keys are rejected with a validation error (only reserved meta keys like `confirm` are stripped before validation). For the JSON Schema of a specific action's `params`, read the MCP resource `gitlab://schema/meta/{tool}/{action}` (replace placeholders with the tool name and the chosen action).

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
next_stepsNoOptional. Suggested follow-up actions or tool calls for the LLM, contextual to the result.
paginationNoPresent on list actions. Use `has_more` and `next_page` to paginate through results.
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Beyond annotations (readOnlyHint, destructiveHint, idempotentHint), description details read-only nature, response format (base64, large responses), error scenarios, and confirms tool always registered. No contradictions.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Well-structured with clear sections, but slightly verbose with error handling details. Front-loaded with purpose and alternatives, then parameter details. Each part contributes meaning, though could trim some repetition.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Covers purpose, usage guidelines, parameters, errors, return format, and links to external schema. Output schema exists and description explains return values adequately. For a tool with nested params and multiple actions, this is very complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%. The description adds significant value by detailing each parameter's type, requirements, and source of values (e.g., path hints, UI navigation). Goes beyond basic schema to aid correct usage.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states 'Download ML model package files from the GitLab Model Registry.' It distinguishes from siblings by explicitly listing 'NOT for' cases and naming alternative tools (gitlab_package, gitlab_release, etc.).

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Provides explicit 'When to use' and 'NOT for' sections, naming sibling tools for clarity. Also mentions prerequisites (Reporter+, Premium/Ultimate plan) and error hints for 404, 403, 400.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/jmrplens/gitlab-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server