Skip to content
Akmelo

Documentation

Getting results

Site records

GET /api/v1/sites/{site}/records returns saved, processed records rather than the source page's HTML. Use /records/{key} for one record and URL-encode the key.

Read run status and logs through /api/v1/runs/{run}. If an agent collected the data, the same central API methods can read results synchronized to the platform.

Manage an export

  1. Read modules with GET /api/v1/exports/modules. Modules marked as planned are not available yet.
  2. Create an export with POST /api/v1/exports using site_id, name, format, and a schema. Omit code to let the platform generate it.
  3. Ask /exports/{id}/suggest-schema for initial fields, then adjust them with PATCH.
  4. Inspect /exports/{id}/preview.
  5. Prepare a file with POST /exports/{id}/run.
  6. Read GET /exports/{id}/runs; prepared files appear in files.

Download with an API key through GET /api/v1/exports/files/{uuid}. Use this versioned path for server integrations even when an interface link uses /api/exports/files/{uuid}. Check the preparation status and file presence: creating an export configuration does not generate a file by itself.

It is often easiest to configure a schema and transformation in the interface first, then inspect GET /exports/{id}. The structure depends on the format. See field mapping and code transformations.

Direct catalog endpoint

For a recipient that only needs the result, issue a separate export token limited to a site, export, and format. The path is /api/export/{site_code}/{export_code}.{format}. JSON, XML, and CSV are supported. This endpoint does not use a management API key.

See export by URL for pagination, cursor, and JSON envelope details. Pagination of prepared data differs from traversing source-site pages; source pagination belongs in a collection scenario.

Deliveries and connections

Create connections through /connections and delivery rules through /exports/{id}/deliveries. For a manual delivery, call POST /deliveries/{id}/run. HTTP 202 means it has entered the queue. Resetting a delivery cursor may send previously delivered records again; agree on duplicate handling with the recipient.

An external database or storage service does not consume Akmelo space for data stored there. Transfer and processing time still count. Using the API does not change these rules.

Product photos

Record schemas support image for a primary photo and gallery for up to 50 photos. A scenario emits a URL or URL list. A production run stores the image files and replaces source values with private references, dimensions, and MIME type. Trial runs do not upload files.

Ready-made Wildberries scenarios provide main_image and gallery. Both can be shown as table columns and opened in the shared zoomable gallery. Site settings select Akmelo storage or an S3 connection belonging to the same workspace. Personal scenarios cannot use another team’s storage. A self-hosted agent stores photos locally by default.