generated: '2026-08-06'
method: derived
source: openapi/a-alpha-bio-atlas-data-product-openapi-original.json
status: candidate
url: null
run_by: none
disclaimer: A-Alpha Bio DOES NOT PUBLISH AN MCP SERVER. This file is an API-Evangelist-derived proposal,
  mapped one-to-one from the nine operations of the Atlas Data Product API. Nothing here is an A-Alpha
  Bio product or endorsement, and no endpoint below exists. It is recorded because the API is read-only,
  small, and describes ML training data — about as MCP-shaped a surface as exists — so the adapter is
  one thin wrapper away.
searched:
  fetched: '2026-08-06'
  probes:
  - url: https://api.atlas.aalphabio.com/mcp
    http_status: 404
  - url: https://api.atlas.aalphabio.com/api/v1/mcp
    http_status: 404
  - url: https://api.atlas.aalphabio.com/api/mcp
    http_status: 404
  - url: https://api.atlas.aalphabio.com/sse
    http_status: 404
  - url: https://mcp.aalphabio.com/mcp
    http_status: 0
    note: DNS does not resolve
  - url: https://api.atlas.aalphabio.com/.well-known/oauth-protected-resource
    http_status: 404
  method: 'POST {"jsonrpc":"2.0","id":1,"method":"tools/list"} with Accept: application/json, text/event-stream'
  result: no MCP surface found on any A-Alpha Bio host
transport_proposed: streamable-http
auth_proposed: Bearer passthrough of the caller's Atlas (AWS Cognito) access token. The three discovery
  tools below work with no token at all, which makes an anonymous read-only server viable for catalogue
  browsing.
tools:
- name: list_datasets
  description: List Atlas Data Blocks. Set include_locked=true to see the public licensable catalogue
    (with defaults, an unauthenticated caller gets an empty list).
  rest: listDatasets
  http: GET /api/v1/datasets
  auth_required: false
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets/get/parameters
- name: get_dataset
  description: Get metadata for one Atlas Data Block, optionally at a pinned version.
  rest: getDataset
  http: GET /api/v1/datasets/{id}
  auth_required: false
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}/get/parameters
- name: get_dataset_datacard
  description: Get the structured Data Card for a Data Block — executive summary, ML guidance (noise/signal,
    recommended splits, data views, column schema) and biology/assay context.
  rest: getDatasetDatacard
  http: GET /api/v1/datasets/{id}/datacard
  auth_required: false
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1datacard/get/parameters
- name: get_dataset_data
  description: Get a pre-signed URL for a Data Block's CSV, optionally row-limited and mode-selected.
  rest: getDatasetData
  http: GET /api/v1/datasets/{id}/data
  auth_required: true
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1data/get/parameters
  note: Use redirect=false so the tool returns a URL as JSON rather than following a 302.
- name: get_dataset_schema
  description: Get the CSV column schema for a Data Block, shaped to unpack into pd.read_csv(url, **schema).
  rest: getDatasetSchema
  http: GET /api/v1/datasets/{id}/schema
  auth_required: true
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1schema/get/parameters
- name: list_dataset_structures
  description: List the structure (.cif) files attached to a Data Block.
  rest: listDatasetStructures
  http: GET /api/v1/datasets/{id}/structures
  auth_required: true
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1structures/get/parameters
- name: download_dataset_structures
  description: Get a pre-signed URL for a zip archive of all of a Data Block's structure files.
  rest: downloadDatasetStructures
  http: GET /api/v1/datasets/{id}/structures/download
  auth_required: true
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1structures~1download/get/parameters
- name: get_dataset_structure
  description: Get the content of one structure file from a Data Block.
  rest: getDatasetStructure
  http: GET /api/v1/datasets/{id}/structures/{filename}
  auth_required: true
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1structures~1{filename}/get/parameters
- name: get_dataset_figure
  description: Get a figure image referenced by a Data Block's Data Card, scoped to a card tab.
  rest: getDatasetFigure
  http: GET /api/v1/datasets/{id}/figures/{tab}/{filename}
  auth_required: true
  inputSchema_source: openapi#/paths/~1api~1v1~1datasets~1{id}~1figures~1{tab}~1{filename}/get/parameters
coverage:
  openapi_operations: 9
  tools_proposed: 9
  mcp_only: 0
  rest_only: 0
deployment:
  mode: none
  verified: derived
  tools: 9
  checked: '2026-08-12'
  source: catalog MCP census