+80
any post, any URL
{"title": "Sunday Meatballs","rating": 4.8,"prep_time_min": 45,"ingredients": [ …12 ]}
200 · schema

Turn any post into typed JSON.

POST postreef.com/v1/extractions

A URL goes in. Your object comes out.

Four steps, and only the first two are yours.

  1. 01

    Paste a link

    Any post or page. A reel, a carousel, a video, an article.

  2. 02

    Pick the fields

    One of 30+ ready-made schemas, your own JSON Schema, or a description we compile into one.

  3. 03

    We fetch and read it

    Rotating proxies through the bot walls, then video, audio, images, transcript and comments read together — or the article text when the URL is a page.

  4. 04

    Get typed JSON

    Every field conforms to the schema you sent. Priced before it runs, refunded if it fails.

You wanted to ship a feature. Not a scraping pipeline.

Pulling structured data out of a post means proxies, bot walls, transcripts and parsing. On every platform, and again every time one of them changes. PostReef is that whole middle, behind one request.

The old way

  • A different scraper for every platform
  • Proxy pools to rent, rotate and replace
  • Headless browsers just to load one page
  • Bot walls and login gates, forever
  • Audio you have to transcribe yourself
  • Parsing whatever comes back into your types

And now you maintain all of that, for every platform, forever.

With PostReef

  • POST the URL and the JSON Schema you want
  • We rotate proxies and get past the bot walls
  • Video, audio, images, metadata, transcript and comments, read together
  • You get back an object that conforms to your schema

One request. Typed JSON. Failed runs refunded in full.

Watch it run on a real post.

Five real posts, replayed. The JSON and every file are yours.

youtube.com/watch?v=B0ZlI8rF8CM
60-Second Video Tips: How to Make Awesome Pickled OnionsAmerica's Test Kitchen · YouTube
Fetching video info
Downloading media
Building transcript
Fetching comments
Running AI extraction
Finishing up

Any format. Same fields.

instagram.com/p/DPwb…reelWorkoutexercises[4]setsrepsinstagram.com/p/DTx7…7 slidesTravel Guidedestinationstops[6]dayssoundguys.com/anker-…articleProduct ReviewproductNameverdictpros[5]

30+ schemas ready · browse them · or describe your own

Or skip the model and just take the files.

Leave the schema out and the run stops after the download. Same fetching, same proxies, same bot walls, no model in the loop. You get the artifacts and you decide what to do with them.

Video

The video file the post is built on.

Audio

The soundtrack on its own, without the video.

Transcript

Captions and subtitles, in the post's own language.

Comments

The comment thread, as JSON.

Images

Carousel slides and photo posts, full size.

Metadata

Title, description, author, counts, thumbnail.

{ "url": "…", "parts": ["video", "transcript"] }

No schema, no AI charge. You pay the base fee and the artifacts you asked for. See the API reference

Priced per run. Here's one.

3-minute video · transcript + comments · Recipe schema

Base feeevery run$0.005

Transcriptflat$0.001

Commentsflat$0.002

AI pass over text3 min$0.009

$0.017170 credits

Quoted firstFailures refundedCache hits free
Credits, not a planTop up when you want, nothing recurring. Your first top-up earns 1,000 bonus credits.
What moves the priceThe artifacts you ask for and how long the video runs. Text is the cheapest thing to read, then audio, then video. Articles and photo posts have no clock, so they're priced flat.
A failed run costs nothingEvery run is quoted before it starts, and one that does not finish is refunded in full, automatically.
Credits keepThey do not expire. Turn on auto-recharge if you want top-ups to happen on their own, or leave it off.

Price your own run

Paste something weird at it

1,000 bonus credits on your first top-up · failed runs refunded in full

Get started →Read the docs
YouTubeTikTokInstagramVimeoTwitchDailymotionPinterest+ any URL