Amyloom Is Live: As-Reported Financial Data Built for AI

Today we are opening Amyloom to beta users. We are putting our financial data API in the hands of developers, analysts, and traders who need what the market currently lacks: structured, as-reported SEC filing data that machines can actually work with.

You can register for beta access today at amyloom.com.

The Problem We Are Solving

Financial data APIs have existed for years. Yet, accuracy remains a persistent problem. We have seen it firsthand: line items that do not match the filings, timestamps that are incorrect, missing amendments, and values that look suspiciously rounded.

The issue runs deeper than accuracy. If you ask, "What did Amazon call its R&D expense last quarter?", you will find it is labeled "Technology and infrastructure" in their filings, not "R&D." If you need to know if a bank is holding unrealized losses in held-to-maturity securities, that detail lives in a specific balance sheet line item with a parenthetical disclosure.

When data quality is inconsistent, these details get lost.

We built Amyloom because we needed this data ourselves. We previously worked on a financial data platform relying on third-party vendors. The quality issues were constant roadblocks: missing data points, inconsistent taxonomies, and errors that required manual reconciliation.

We spent too many hours extracting numbers from PDFs and building scrapers that broke every time a company changed their filing format. We decided to build the infrastructure we wished existed.

Data Coverage & Capabilities

Amyloom provides access to every line item from 10-K, 10-Q, 10-K/A, 10-Q/A, and S-1 filings, exactly as companies report them. We use AI to extract the data, but every extraction is validated by our team before production. The result is clean, structured JSON that preserves the original taxonomy, labels, and values.

Coverage and Features

• 6,500+ US stocks with 20+ years of historical data.
• Near real-time extraction: Data is available minutes after filings hit the SEC website.
• Full traceability: Every number links back to the specific line item in the original filing.
• Amendment tracking: We capture both original and amended filings with SEC acceptance timestamps.
• S-1 access: Access to pre-IPO financials and risk factors.

Why Amendments Matter

When a company restates prior period financials, you need to know what changed, when it changed, and the magnitude of the correction. We capture both the original filing and the amended version, timestamped with SEC acceptance times. This allows you to update historical models accurately without erasing the past.

Use Cases

Investment Teams & Asset Managers

Build AI agents that understand what companies actually report. Automate quality of earnings analysis, screen for buried expenses in "other" line items, and benchmark segment-level unit economics.

Fintech Developers

Build features that differentiate your product. Create native AI assistants that cite their sources, deep screening tools that filter on segment-level margins, or risk monitors that flag specific balance sheet disclosures.

Quants and Traders

Ensure point-in-time accuracy. SEC acceptance timestamps prevent look-ahead bias in backtests. Track amendments automatically and build factor models on granular line items using exact integers, not rounded aggregates.

As-Reported vs. Standardized Data

The industry standardized financial data because human analysts needed comparability. If you compare Microsoft and Apple, you want their income statements to use the same labels.

AI models do not need this. LLMs understand that "Cost of revenue" and "Cost of sales" mean the same thing. What they struggle with is hallucination and source attribution.

When an AI model uses normalized data, it cannot link back to the source document. When it uses as-reported data, every answer can cite the exact SEC filing, section, and line item.

We are not replacing existing data vendors. We are building the infrastructure layer that makes SEC filing data work reliably with LLMs and agents.

Beta Access

January 2026: Website launch, beta registration opens.

Early February 2026: Datasets ready, beta access begins.

Q2 2026: Full API launch with real-time extraction.

Beta users benefits:

• Free access to complete datasets during the beta.
• Permanent discount on selected plans at launch.
• Priority support as we refine the product.

Our Commitment

We are starting with US equities because the data quality and regulatory framework are strong. Over time, we will expand coverage and release open-source tools. But our core principle remains: every number must be traceable to its source, validated, and usable for developers.

Financial markets run on information. For too long, that information has been locked in PDFs or buried in databases with quality issues. We are changing that.

Join the beta at amyloom.com.