Thirteen models.
One fair value.
Every trading day.
CirclFi computes 76,401 valuations a day across 5,877 US equities — straight from SEC filings, through thirteen independent institutional models — and shows you exactly where price and value disagree.
Showing the 4 valuations that are free on every stock page. The remaining models are part of the terminal.
CirclFi is an independent equity valuation engine. It runs 13 institutional valuation models — Bayesian DCF, Greenwald Earnings Power Value, McKinsey EROIC spread and ten more — on 5,877 US-listed stocks every trading day, computing 76,401 valuations per run from SEC EDGAR filings, XBRL financials, FRED macroeconomic series and GDELT news sentiment. Each stock gets a consensus intrinsic value, a 32-signal quality score and a value-trap flag.
How do thirteen valuation models become one fair value?
Each valuation model attacks the same company from a different angle — discounted cash flow, relative, scenario, option-based. The CUCE ensemble then weights all twelve outputs by correlation-unbiased certainty equivalence to produce the consensus intrinsic value in the middle of the dial. Where the models disagree most is where the market is most likely mispricing the stock.
| # | Model | Dispersion vs price | Value |
|---|---|---|---|
| 01 | Bayesian DCF | terminal | |
| 02 | Earnings Power Value | terminal | |
| 03 | EROIC Spread | terminal | |
| 04 | First Chicago | terminal | |
| 05 | Markov DDM | terminal | |
| 06 | ML-RIV | terminal | |
| 07 | Dynamic NAV | terminal | |
| 08 | PWERM | terminal | |
| 09 | Regime Cross-Sectional | terminal | |
| 10 | Sentiment SOTP | terminal | |
| 11 | CUCE Ensemble | terminal | |
| 12 | FTNN Topology | terminal | |
| 13 | RCMH-DCF | terminal |
How does CirclFi turn SEC filings into a fair value?
Ingest & scan
Compute & value
Signal & act
Which 13 valuation models does CirclFi run?
III — THE ENGINE| # | Model | Class | What it does | Best for |
|---|---|---|---|---|
| 01 | Bayesian DCF | Intrinsic | 10,000 Monte Carlo + Merton jump-diffusion | Stable FCF companies |
| 02 | Earnings Power Value | Intrinsic | Greenwald Franchise Value Trilogy + RAV | Mature earners, Moat companies |
| 03 | Markov DDM | Intrinsic | Regime-switching + total shareholder yield | Dividend payers, Buyback-heavy (AAPL/GOOG) |
| 04 | Dynamic NAV | Asset-Based | Sector-specific recovery rates + hidden liabilities | REITs, Banks, Insurance, Resource companies |
| 05 | EROIC Spread | Intrinsic | McKinsey Value framework + CAP fade + R&D capitalization | Moat companies, High-ROIC businesses |
| 06 | ML-RIV | Intrinsic | 5-factor DuPont ROE decomposition + persistence modeling | Banks, Financials, Insurance |
| 07 | First Chicago | Scenario | Cyclicality-adjusted probability weights + sector benchmarks | Growth stocks, Biotech, Cyclicals, Turnarounds |
| 08 | PWERM | Option-Based | Merton structural model + stochastic volatility | Distressed companies, M&A targets, High-leverage |
| 09 | Regime Cross-Sectional | Relative | 6-indicator macro regime + sector-adjusted multiples | All sectors with sufficient peers |
| 10 | Sentiment SOTP | Hybrid | EDGAR segment data + GDELT news sentiment | Conglomerates, Diversified, Multi-segment |
| 11 | CUCE Ensemble | Ensemble | Correlation-adjusted inverse-variance weighting | All sectors (meta-model) |
| 12 | FTNN Topology | Relative | 6D Gaussian kernel financial similarity | Tech, Healthcare, Consumer, any sector |
| 13 | RCMH-DCF | Intrinsic | 4-regime parallel DCF + FRED macro integration | Rate-sensitive, Utilities, REITs, Cyclicals |
What research is the valuation engine built on?
IV — FOUNDATIONSMerton–Black–Scholes
The Greenwald trilogy
McKinsey economic profit
Bayesian Monte Carlo
Every formula, every derivation, every data source — documented in our open methodology. Read the full methodology →
Why does a fair value need a quality score beside it?
32-signal QOC score
Value-trap detection
Live terminal, not a PDF
Whole-market coverage
Can my AI assistant use the CirclFi engine?
The CirclFi MCP server turns Claude Desktop — or any MCP-compatible agent — into an institutional analyst with direct access to all thirteen models across 5,877 stocks.
BLOOMBERG TERMINAL
FACTSET
CIRCLFI
How much does CirclFi cost?
VII — PRICINGWhat CirclFi is not
What is CirclFi and how does it work?
What is CirclFi?
CirclFi is an independent equity valuation engine. It runs 13 institutional models on 5,877+ US-listed companies every trading day — Bayesian DCF with 10,000 Monte Carlo paths, Greenwald Earnings Power Value, McKinsey EROIC spread and ten more — and publishes every output in a live terminal. Each stock gets a consensus intrinsic value, a quality score and a value-trap flag.
How many stocks does CirclFi cover?
CirclFi tracks 5,877+ US equities across all major exchanges — every company that files with SEC EDGAR, from mega-caps to micro-caps. Valuations are recalculated daily after the US market closes, producing over 76,401 individual model valuations per engine run.
Which 13 valuation models does CirclFi run?
Bayesian DCF, Earnings Power Value (EPV), EROIC Spread, First Chicago, Markov DDM, ML Residual Income (ML-RIV), Dynamic NAV, PWERM, Regime Cross-Sectional, Sentiment SOTP, CUCE Ensemble, FTNN Topology and RCMH-DCF. Each attacks intrinsic value from a different angle — discounted cash flow, relative, scenario, option-based, asset-based and ensemble — so the spread between them is itself a signal.
Where does the data come from?
SEC EDGAR filings and 700+ standardised XBRL financial tags, FRED macroeconomic series (risk-free rates, VIX, GDP growth, yield curve, inflation), GDELT global news sentiment and public market price data. Nothing is hand-entered and nothing is borrowed from another newsletter. Every figure traces back to a public filing.
How much does CirclFi cost?
Standard Access is $39 a month or $299 a year, with the full 13-model engine, the live web terminal, QOC and value-trap scores and the methodology whitepaper. Institutional Access is $399 a month or $3,499 a year and adds five team seats plus a dedicated API connection. Free visitors see a subset of models on every stock page.
Is CirclFi financial advice?
No. CirclFi is an algorithmic valuation tool for educational and research purposes. There are no price predictions and no buy or sell signals. It computes what 13 mathematical models imply about a stock's intrinsic value from public SEC filings — what you do with that analysis is your own decision.
Do the AI agents do the valuation?
No — the mathematics is deterministic and reproducible from the filings. Our 327 agents read daily news against twenty years of historical events to adjust forward-looking assumptions, always inside a strictly conservative baseline. The models themselves are unchanged by them.
Can my AI assistant use CirclFi?
Yes. The CirclFi MCP server turns Claude Desktop or any MCP-compatible agent into an analyst with direct access to the engine — screen 5,877 stocks in milliseconds and pull all thirteen valuations plus the QOC score. It runs as a pure edge-compute API on Cloudflare with no local database.
How is CirclFi different from a Bloomberg Terminal?
A terminal sells you data and leaves the modelling to you. CirclFi ships the finished models — the consensus fair value, the quality score and the value-trap flag — at roughly two percent of the cost. Bloomberg runs about $2,000 a month and FactSet about $1,000; CirclFi starts at $39.
Full model documentation: CirclFi methodology · Pricing detail: plans and costs
The market misprices something every single day.
The only question is whether you're the one who sees it. 13 models · 5,877 stocks · recomputed before tomorrow's open.
Buy with Gumroad