Financial data that tells the truth
We built Valuein because the financial data market is broken. Expensive Bloomberg terminals. Survivorship-biased datasets. Stale Compustat snapshots that don't tell you what investors actually knew at each point in time. We built the alternative.
105M+
Standardized Financial Facts
12,000+
Companies Covered
1994
History Starts
8
Parquet Tables Per Tier
The problem we're solving
Most financial data products fail in one of three ways: they're too expensive for individual analysts and small funds, they carry survivorship bias that inflates backtest returns, or they don't tell you when the data became available — making point-in-time analysis impossible.
The SEC has been collecting structured financial data since the 1990s. Every public company files quarterly and annually. That data is public. But parsing, normalizing, and standardizing 105 million XBRL facts across 12,000 companies and 30 years is not a weekend project.
We did that work so you don't have to. Every fact carries its knowledge_at timestamp — the moment it entered our system from EDGAR. Every amendment is tracked. Every delisted company is preserved. The data pipeline runs daily.
What we stand for
Six principles that guide every product decision.
Point-in-Time Accuracy
Every data point carries the timestamp of when it was known — not when it was filed. Your backtests see exactly what investors saw, before restatements, before revisions.
Zero Survivorship Bias
Delisted, bankrupt, acquired — they're all here. A strategy that only works on survivors isn't a strategy. We include every entity that ever filed with the SEC.
Standardized Across Decades
AAPL Q1 2001 and Q1 2024 use the same column names. We normalize XBRL tags, handle fiscal year mismatches, and map amendments to their originals so your queries always work.
Built for Speed
Parquet over R2 means you download only what you need and query with DuckDB at sub-second speed. No ORM, no polling, no pagination — column-oriented data the way quants want it.
Amendment Tracking
When a company restates earnings, we record both the original reported value and the corrected value. You can filter by knowledge_at to reconstruct any historical view.
Open Source Distribution
The valuein Python SDK is open source. No black-box query builders, no vendor lock-in. You write SQL over Parquet — we provide the data engine and the distribution.
Three ways to access the data
Whether you're a Python quant, an Excel analyst, or building a data product, Valuein meets you where you work.
- Python SDK →
pip install valuein — from install to a DuckDB-backed DataFrame in 60 seconds. SQL-native, zero black-box abstractions.
- Excel Power Query →
Connect your spreadsheet to live SEC data. Models refresh automatically when filings drop. No VBA. No copy-paste.
- MCP Server →
Connect Claude, Cursor, or any MCP-compatible AI assistant directly to the dataset. Query 105M+ financial facts with natural language — no SQL required.
Data pipeline
SEC EDGAR filing drops
10-K, 10-Q, 8-K, 20-F
Ingest & normalize XBRL
105M+ facts, standardized concepts
Point-in-time indexing
knowledge_at timestamp assigned
Amendment reconciliation
10-K/A and 10-Q/A tracked
Parquet export to R2
Per-tier buckets, updated daily
Available via SDK / API
< 24h after EDGAR availability
SEC filing coverage, start to finish
10-K, 10-Q, 8-K, 20-F — every major form type ingested and standardized. Click any to see what's available and how to query it.
Ready to ditch survivorship bias?
Free tier includes 100 API calls/day. No credit card required.