Code Ocean MCP Server

Search for capsules with filtering, sorting, and pagination options. Compact results: {items: [{id, n, s, d, t}], has_more, next_token, item_count}.

Item fields: id=id, n=name, s=slug, d=description (truncated), t=tags (limited). Pagination: item_count returns the number of items in the current page. Use next_token for additional pages when has_more=true. Set include_field_names=true to add field_names with full labels. Use get_capsule(id) if full details needed.

Search for pipelines with filtering, sorting, and pagination options. Compact results: {items: [{id, n, s, d, t}], has_more, next_token, item_count}.

Item fields: id=id, n=name, s=slug, d=description (truncated), t=tags (limited). Pagination: item_count returns the number of items in the current page. Use next_token for additional pages when has_more=true. Set include_field_names=true to add field_names with full labels. Use get_capsule(id) if full details needed.

Retrieve metadata for a specific capsule by its ID.Use only to fetch metadata for a known capsule ID. Do not use for searching.

Get all computations associated with a specific capsule.

Attach one or more data assets to a capsule with optional mount paths. Use when the capsule has no running cloud workstations. For active cloud workstation sessions use attach_computation_data_assets instead. Accepts a list of parameter objects (e.g. [{'id': '...'}]), not just a list of IDs.

Detach one or more data assets from a capsule by their IDs. Use when the capsule has no running cloud workstations. For active cloud workstation sessions use detach_computation_data_assets instead.

Retrieve app panel information for a specific capsule by its ID.

Search for data assets with filtering, sorting, and pagination options. Compact results: {items: [{id, n, d, t}], has_more, next_token, item_count}.

Item fields: id=id, n=name, d=description (truncated), t=tags (limited). Pagination: item_count returns the number of items in the current page. Use next_token for additional pages when has_more=true. Set include_field_names=true to add field_names with full labels. Use get_data_asset(id) if full details needed.

Get full details for a data asset by ID. Use after compact search to retrieve complete metadata.

Generate view and download URLs for a specific file from an internal data asset.Call only when the data asset is already created and in a ready state. If the asset may not yet be ready, first use wait_until_ready to poll until readiness, then retrieve the download URL.

Use when you want to read the content of a file from a data asset

List files and folders within an internal data asset at the specified path. Empty path retrieves root level contents.

Update metadata for a data asset including name, description, tags, mount, and custom metadata.

Supports updating various metadata types:

• Basic metadata: name (display name), description (free text description)

• Organization: tags (keywords for searching), mount (default mount folder path)

• Custom metadata: admin-defined custom fields with user-set values according to deployment configuration (string, number, or date fields in unix epoch format)

Poll a data asset until it reaches 'Ready' or 'Failed' state with configurable timing.

Args: data_asset: The data asset object to monitor polling_interval: Time between status checks in seconds (minimum 5 seconds) timeout: Maximum time to wait in seconds, or None for no timeout

Returns: Updated data asset object once ready or failed

Raises: ValueError: If polling_interval < 5 or timeout constraints are violated TimeoutError: If data asset doesn't become ready within timeout period Poll until the specified data asset becomes ready before performing further operations (e.g., downloading files). You can set polling_interval and optional timeout.

Create a new data asset from various sources including S3 buckets, computation results, or combined assets.

Data assets are versioned, immutable collections of files that serve as inputs or outputs for computational workflows in Code Ocean. Internal data assets store files within Code Ocean's infrastructure, while external data assets reference files in external storage (S3/GCP) without copying.

Supports creating data assets from AWS S3, GCP Cloud Storage, computation results, or combining existing data assets. Returns confirmation of creation request validity, not success, as creation takes time. Use wait_until_ready() to monitor creation progress.

You can link to the created data assets with the 'data_asset_id' with the pattern: https://codeocean.example.com with /data-assets/<data_asset_id>.

Retrieve metadata and status information for a specific computation by its ID.

Execute a capsule or pipeline with specified parameters and data assets.

For capsule execution: Set run_params.capsule_id and optionally provide data_assets, parameters, or named_parameters.

For pipeline execution: Set run_params.pipeline_id and optionally provide data_assets, processes (with process-specific parameters), and nextflow_profile configuration.

Typical workflow: 1) run_capsule() to start execution 2) wait_until_completed() to monitor progress 3) list_computation_results() and get_result_file_urls() to retrieve outputs.

Poll a computation until it reaches 'Completed' or 'Failed' state with configurable timing.

Args: computation: The computation object to monitor polling_interval: Time between status checks in seconds (minimum 5 seconds) timeout: Maximum time to wait in seconds, or None for no timeout

Returns: Updated computation object once completed or failed

Raises: ValueError: If polling_interval < 5 or timeout constraints are violated TimeoutError: If computation doesn't complete within the timeout period

List result files and folders generated by a computation at the specified path. Empty path retrieves the /results root folder. computation_id is required as string

Generate view and download URLs for a specific result file from a computation.

Use when you want to read the content of a file from a computation

Rename an existing computation with a new display name.

Delete a computation and stop it if currently running.

Attach one or more data assets to a cloud workstation session computation. Accepts a list of parameter objects (e.g. [{'id': '...'}]). Use for cloud workstation sessions.

Detach one or more data assets from a cloud workstation session computation by their IDs.

Retrieve the Code Ocean deployment's custom metadata schema.