Valuein

Institutional-grade U.S. fundamentals data for quants, analysts, and AI agents.

Valuein turns the complexity of SEC EDGAR bulk filings into clean, structured, developer-friendly datasets. We bridge the gap between raw 10-Ks and quantitative research, so you spend your time building models — not cleaning data.

Explore the data · Pricing · Discord · X / Twitter · Public repo

Most financial-data vendors deliver restated, retroactively adjusted numbers. That's fine — until your backtest accidentally sees a 2024 restatement while simulating a 2023 trade. Valuein preserves filings exactly as they were reported and exposes the same data through every channel you already use.

The same dataset, delivered five ways so it lands where you already work. One Stripe-issued token unlocks every channel — no per-channel billing.

fromvaluein_sdkimportValueinClientwithValueinClient() asclient:
df=client.query(""" SELECT symbol, name, sector FROM "references" WHERE is_sp500 = TRUE AND is_active = TRUE ORDER BY name LIMIT 10 """)
print(df)

That's a real query against the live S&P 500 sample. Add VALUEIN_API_KEY only when you need full universe or full history.

Want an AI agent to query for you instead? Add this to claude_desktop_config.json:

{
"mcpServers": {
"valuein": {
"url": "https://mcp.valuein.biz/mcp",
"headers": { "Authorization": "Bearer YOUR_VALUEIN_API_KEY" }
}
}
}

Same URL works for any MCP-capable client — Cursor, Codex, custom LangGraph or CrewAI agents.

The Python SDK ships from PyPI, the MCP server runs as a Cloudflare Worker, the data pipeline ingests SEC EDGAR daily, and the website is built on Next.js + Cloudflare. Source for those lives in private repos until we open-source them — follow the public repo for releases.

SEC EDGAR → Pipeline (Python + Pydantic) → (Parquet Files)
│
┌──────────────┼──────────────┐
▼ ▼
Python SDK MCP server (DuckDB) (mcp.valuein.biz) │ │
└──────────────┼──────────────┘
▼
One Bearer token, every channel
(Stripe-issued, validated at the edge)

The pipeline ingests EDGAR within ~60 seconds of acceptance, normalizes 11,966 raw XBRL tags into ~150 canonical concepts via a deterministic waterfall, writes Parquet snapshots to R2, and serves them through a Cloudflare Worker gateway with token-aware tier routing — Pro and Enterprise tokens see the full universe, Free tokens see the S&P 500, anonymous sees the last five years of S&P 500.

• Quantitative researchers — point-in-time data and survivorship-bias-free coverage for rigorous backtesting and algorithmic strategy development
• Event-driven funds — 8-K coverage captures intra-quarter material events (CEO departures, bankruptcies, M&A) for latency-sensitive strategies
• Financial analysts — as-reported filings with deep history for fundamental due diligence and company research
• Portfolio managers — sector- and factor-relative screens with consistent, audited inputs
• Data engineers — Parquet-native, cloud-distributed datasets that plug straight into existing data infrastructure
• AI / agent builders — natural-language access to the same data through the MCP server, no SDK required
• Academic researchers — the complete historical universe for empirical studies without selection bias

We're building a financial operating system — the canonical place AI agents and humans go for U.S. fundamentals.

• ✅ Point-in-time, survivorship-bias-free Parquet on R2 (1994–present)
• ✅ Python SDK with 44 pre-built SQL templates and a multi-factor alpha framework
• ✅ MCP server with 14 live tools + 8 analyst SOP prompts
• ✅ Excel template with Power Query and 8 pre-configured sheets
• 🔄 Semantic search over filing text (Risk Factors, MD&A, Business, Legal, Controls) — Vectorize backfill in progress
• 🔄 Real-time 8-K push via webhook (Custom tier)
• 🛣️ Programmatic ticker pages with JSON-LD for AEO discovery
• 🛣️ Open-source notebook templates (factor screens, DCF, earnings momentum)

We welcome examples, notebook improvements, query recipes, doc fixes, and data edge-case reports.

• 🐛 File an issue — data quality, feature request, outage, or general question
• 💬 Discuss in Discord
• 📜 Read CONTRIBUTING.md and our Code of Conduct

We're actively looking for contributors to our SDK examples and community-led valuation models. Lifetime access tokens are available for high-quality contributions — see Discord for details.