Any source shape
Tube sites, gallery networks, member areas with logins, RSS, JSON APIs, or a donor whose HTML changes every other month. Cookie and session handling included.
$ parse
A site that fills itself
An empty catalog earns nothing, and hand-posting does not scale past a few hundred items. We write the parser for your specific sources, run it until it is boring, and hand you the code.
Python or C#, running on your server or ours · code and README are yours
$ ./ingest.py --source donor --since 24h --limit 200
found 187 new releases
[ ok ] metadata parsed · models, studio, tags
[ ok ] covers + 12 screens each → webp/avif
[ ok ] watermark removed on 41 clips
[ ok ] uploaded to host · links attached
[warn] 9 skipped · duplicate of existing post
[ ok ] 178 published via content api
next run in 24h · telegram report sent
capabilities
What a parser can actually do
Real metadata
Models, studios, categories, tags, duration, quality, file size and release date — normalized to your taxonomy instead of dumped into one tag soup.
Media pipeline
Covers, screenshot grids, hover preview clips, re-encoding, thumbnails at the right sizes, WebP and AVIF companions generated on the way in.
Watermark removal
Detection and inpainting for burned-in logos, including ones that drift or change position between frames — with a quality gate that skips clips it cannot do cleanly.
File hosts
Upload, remote-upload, link normalization, dead-link sweeps against host APIs, and re-hosting when a provider dies or bans your account.
Publishing
Straight into WordPress or the engine through a versioned content API — scheduled, rate-limited, with duplicate detection and a Telegram report per run.
Written for the day it breaks
Donors change their markup, hosts change their API, quotas appear out of nowhere. Parsers get state files, resumable runs, dry-run modes and undo passes — so a bad night costs you a rerun, not a cleanup of 4,000 broken posts.
Legal is your decision
We build tooling; what you ingest and whether you have the rights to it is yours to decide. We will not touch sources that are obviously illegal, and we will say so plainly rather than quietly billing you.
process
From donor to published
Look at the source
We check how hard it is: markup stability, rate limits, logins, anti-bot, media layout. You get a yes/no and a price before anything is written.
Prototype on 50 items
A small batch lands on staging. You look at how posts turn out and we adjust the mapping until they look native.
Backfill
The historical catalog goes in at a rate your server tolerates, with progress you can watch and stop at any point.
Keep it running
Scheduled runs with reports. Maintenance when a donor changes is a small monthly fee — or you take the code and do it yourself.
Send the donor URL
We will tell you whether it can be parsed, how cleanly, and what it costs — usually the same day.