Export VK Post Images, Videos, Polls, and Attachment Metadata

VK wall text alone can miss the substance of a post. A public update may be a gallery, short video, poll, document, or repost whose caption is only a pointer. The VK Posts Scraper exports a media array and a compact mediaTypes list alongside every saved post.

Collect media-bearing posts

Submit the public communities or profiles and keep a bounded recent window.

{
  "targets": ["vk", "another_public_community"],
  "publishedAfter": "2026-07-01",
  "maxPostsPerTarget": 200,
  "maxTotalPosts": 400,
  "includeReposts": true
}

A photo attachment can include its stable media ID, owner, album, publication time, and the largest public image URL exposed in the page payload. Video records can include title, description, duration, public view count, and publication time. Polls preserve their question, vote count, and answer choices when present.

Flatten attachments into a media table

Keep the post table and create a second attachment table linked by owner and post IDs.

media_rows = []
for post in posts:
    for position, media in enumerate(post.get("media", [])):
        media_rows.append(
            {
                "owner_id": post["ownerId"],
                "post_id": post["postId"],
                "position": position,
                "post_url": post["sourceUrl"],
                "posted_at": post["postedAt"],
                "type": media.get("type"),
                "media_id": media.get("id"),
                "title": media.get("title"),
                "image_url": media.get("imageUrl"),
                "duration_seconds": media.get("durationSeconds"),
                "media_views": media.get("viewsCount"),
            }
        )

Do not assume a post has only one image or video. Preserve attachment order and the original nested object so newly relevant fields do not require another scrape.

Preserve repost provenance

When repost is present, the outer post and source post are different objects. The outer post records who amplified the content and when; the nested source records the original owner, post ID, text, date, and source URL available in copy history. Retaining both prevents an analysis from crediting every reused asset to the republishing account.

Treat URLs as references, not permanent files

Public media URLs can expire, redirect, or become inaccessible after moderation or privacy changes. The Actor does not promise a permanent CDN archive. Store the source URL, attachment IDs, and scrape timestamp even when downloading is allowed. Follow copyright, data-protection, and platform requirements; public visibility does not remove ownership or grant unrestricted reuse.

For research, consider storing hashes and derived labels instead of redistributing the media. For monitoring, alert on new attachment IDs and retain only the minimum evidence required by the workflow.