3spread

API

Uncompromising access to fundamental data

Standardized and ready for running research and training models with 500K+ tracked filers and billions of data points.

Get startedRead the docs

Programmatic access to the entire 3spread dataset

The same standardized fundamentals, ownership, and fund data behind the terminal, exposed as a clean REST API for your own systems. Point your code at one base URL and pull any filer, any period, and any filing family, without scraping EDGAR or wrangling raw XBRL yourself.

It is built for technical teams: predictable endpoints, typed responses, and a schema that stays the same as your coverage grows from one ticker to the whole market.

Python and a curl request calling the 3spread API

Designed to drop into your stack

There is nothing bespoke to learn before your first call. Everything speaks the conventions your tools already expect.

  • Standard HTTP, JSON, and one sk_live_ key, plus an official Python SDK over the same /v1 endpoints
  • One schema across every filer, so a query written once works everywhere
  • Point in time and deterministic: as-reported figures stay separate from restatements, and the same filing returns the same output every run
  • Verified: every statement reconciles to the filer's own reported totals before it is served
  • Zero generative AI: only the filer's facts and our arithmetic, nothing paraphrased or inferred
The 3spread stack: the API over an analytics pipeline and application backend, with auth, a web GUI, and users

Live within seconds of the filing

As the SEC accepts a filing, it is parsed, standardized, and served within seconds. Your monitoring and your models act on it while it still moves the market, not after an overnight batch.

  • A per-family changefeed streams created and amended events off a cursor, so incremental sync is one poll, not a full re-scan
  • A freshness watermark reports the latest accepted time per family, so you can tell 'caught up' from 'stale' programmatically
  • Amendments land as new, linked versions, so you see the change without losing the original
  • The same pipeline serves live data and history going back to 2012
A live status view of filings being parsed, standardized, and served within seconds of SEC acceptance

What teams build on it

Backtesting and factor models

Point-in-time panels mean models train on what was known at each date, never on a restatement that only landed later.

AI agents and automation

Structured, source-linked data an agent can cite keeps automated research grounded in the filing it came from.

Data pipelines and warehouses

One REST API and clean JSON load straight into your pipeline, with no per-filer cleanup layer in between.

Research at scale

Screen the entire universe or pull a single section of a single filing, both in one call.

Every dataset behind one API

46 endpoints across every filing family, all on the same authentication and the same response shape.

Browse the datasets

Start building

Where you go next depends on what you're after.

Read the docs

Every endpoint, parameter, and response schema.

Open docs

Get started

Full API access and every dataset, for personal research and analysis.

Sign up

Browse the datasets

Every filing family, and what is in each.

See datasets