Quickstart
Up and running in 5 minutes.
No account needed to start. Hit a public dataset right now, then create your own when you are ready.
Hit a public dataset. No account required.
Every public dataset on Quorel is accessible to anyone. No API key, no signup. Just a GET request.
Request
GET https://quorel.vercel.app/api/1/hacker-news-top/active/
?format=json
&limit=5
&sort=score:descResponse
[
{
"title": "Show HN: I built a source of truth for the web",
"score": 412,
"comments": 187,
"url": "https://news.ycombinator.com/item?id=..."
},
...
]That is it. No setup. Browse all public datasets in the marketplace and swap the slug for any of them.
API URL structure
https://quorel.vercel.app/api/{dataset_id}/{name-slug}/{version}/dataset_id
The numeric ID of the dataset.
name-slug
The dataset name, lowercased and hyphenated. Must match exactly.
version
active (the pinned version), latest (the most recent), or v1, v2, v3 etc.
Query parameters
Every endpoint accepts the same set of params. Stack as many as you need.
Output format. json, jsonl, csv, tsv, xml, parquet.
Max number of entities to return.
Skip the first N entities.
Sort by a field. sort=score:desc or sort=name:asc.
Full-text search across all fields. Comma-separated for OR. Use keywords_mode=and for AND.
Exact match on a field. filter=category:finance. Repeatable.
Partial match on a field. filter_contains=title:startup. Repeatable.
Return only these fields. Dot-notation supported. keep_field=title,score.
Remove these fields from the response. drop_field=_source.
Remove duplicate entities. Use dedup_key=field1,field2 to dedup on specific fields.
Strip null and empty string values recursively.
Flatten nested objects into dot-notation keys. Required automatically for csv, tsv, parquet.
Return N randomly sampled entities.
Return only the count of matched entities, not the entities themselves.
Return the alt version of the dataset if one exists.
Include the _source field (the origin URL of each entity). Set to false to strip it.
Pretty-print JSON output. Set to false for compact output.
Rate limits
Public (no auth)
100 requests per minute, per IP
Private (Bearer token)
300 requests per minute, per IP
Rate limit headers are included in every response: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset.
Create your own dataset.
Free account, no credit card. One dataset, up to 20 URLs.
Sign up
Create a free account at https://quorel.vercel.app/auth. No credit card required.
Paste URLs or describe your intent
Import URLs directly, or type a plain-English description of what you want and Quorel will discover the sources for you using web search.
Define your schema
Tell Quorel which fields you want and what they mean in plain English. For example: title (string) -- the title of the article, score (number) -- the upvote count.
Wait for processing
Quorel crawls your URLs, runs AI extraction against your schema, and versions the result. You will get a notification when it is ready. Usually takes a few minutes.
Hit your endpoint
Your dataset is live. Find your dataset ID and name slug in the dashboard, then hit the API exactly like you did in Path 1.
Accessing a private dataset
Private datasets require a Bearer token. Find and copy your private key in your dataset inside your dashboard.
GET https://quorel.vercel.app/api/42/my-dataset/active/ Authorization: Bearer your_private_key