Use the data

Access and download data

On this page

Public records and their available files can be accessed without an account. Use the website to explore individual entries and the API to retrieve records programmatically.

Download individual files

Open an entry or model and select an available file's download link. Files may be stored by The Dynamic PDB or linked from an external source. Available content varies by entry and may include experimental data, coordinates, and supporting files.

Use the API

The API provides public records without authentication. Responses use JSON format, with records inside a data field. Scripts and LLM agents should use these structured records to retrieve metadata and file references. The API reference describes every endpoint and request format.

For reproducible analysis, retain the entry and model identifiers, file checksums where available, and the date of retrieval. Public records may change when a newer revision is approved.

1. Find an entry by PDB ID

curl -fsS \
  -H 'Accept: application/vnd.api+json' \
  'https://dynamicpdb.com/api/v1/entries?pdb_id=7C24'

Use the returned data[].id for subsequent requests. The entry ID in The Dynamic PDB is distinct from its external PDB reference.

2. List the entry's models

curl -fsS \
  -H 'Accept: application/vnd.api+json' \
  'https://dynamicpdb.com/api/v1/entries/dpdb_ube2dzua/models?limit=100'

Each model includes its identifier, metadata, and reported metrics. Per-residue data (residue_data) is returned only when you request a single model, at /v1/entries/{entry_id}/models/{model_id}.

3. Find a model's coordinate file

curl -fsS \
  -H 'Accept: application/vnd.api+json' \
  'https://dynamicpdb.com/api/v1/entries/dpdb_ube2dzua/models/dpdb_ube2dzua_m_002/artifacts?levels=L2&limit=100'

The API calls file records artifacts. Each record can include its name, format, checksum, size, and file location in attributes.uri.

L2 results can include both coordinates and supporting files, such as refinement logs. Select the coordinate record using its type and format, or match its ID to the model's primary_artifact_id.

4. Download the selected file

Copy the selected record's HTTP or HTTPS attributes.uri into this command:

curl -fL 'REPLACE_WITH_FILE_URI' -o model.pdb

Use an output extension that matches the file format.

Retrieve larger collections

Bulk download is not currently implemented. Scripts can enumerate records through the API and download available files individually.

List endpoints return at most 100 records per request; set limit=100 to get the most per page.

Every list is paged by cursor: entries, models, file lists, similar entries and revisions. When more records exist, the response includes links.next, the full URL of the next page with the same filters. Request it, and repeat until a response has no links. To build the URL yourself, pass the response's meta.next_cursor as the cursor parameter. The offset parameter still works but is deprecated: deep offsets are slower, and records can be skipped or repeated while the data changes.

Retrieve entry-level and model-level files separately. A model's file list also returns the model's runs in meta on every page.

The OpenAPI specification is a machine-readable description of the API, suitable for scripts and LLM agents.

Before reusing data, check the applicable license and attribution requirements, including those of externally referenced sources. Data originating from The Dynamic PDB and Prism are freely available for community use. See Reuse, acknowledgment, and feedback.