Projects Financial Spreading + Comps Agent
Structured financials and public comps from any 10-K in 45 seconds.
Replaces 6–7 hours of manual spreading and CapIQ work. Normalized to millions USD, LTM plus 3 fiscal years. EV/EBITDA and revenue multiples for up to 10 comparable public companies — exported to CSV.
SEC EDGAR filings · Confidence-scored line by line · Audit-ready citations
- Full pipeline
- 45 sec
- Of manual work replaced
- 6–7 hrs
- Public comps
- 10
- Data source
- SEC EDGAR
Before / after
From a working day to under a minute.
| Task | Manual process | Agent |
|---|---|---|
| Spreading a 10-K | 2–4 hours | ~15 seconds |
| Building a comps table | 3–4 hours | ~30 seconds |
| CapIQ subscription required | Yes | No |
| Error-prone transcription | Yes | No |
| Audit-ready citation trail | No | Yes |
Processing pipeline
Four steps, no reformatting.
- 01
PDF upload
Drop your 10-K PDF (up to 50MB). Accepts annual filings from any public company. No preprocessing or formatting required.
- 02
Claude extraction
Claude Sonnet 4.6 isolates the financial statement sections and spreads IS / BS / CF to structured JSON at temperature 0. A second pass validates sign conventions, label alignment, and unit consistency.
- 03
EDGAR + yfinance
The SEC EDGAR XBRL API identifies SIC-matched peer companies. yfinance pulls market capitalization, enterprise value, and trailing price data for each comparable.
- 04
CSV export
A two-section spreadsheet: financial spread plus comps table, normalized in millions USD. Drops directly into your existing Excel model.
Sample output
What lands in your spreadsheet.
A fictional company, shown for illustration. Real output is built from the filing you upload.
Spread financials
Acme Corp (ACME) · 10-K · FY2024 · USD millions
| Metric | LTM | FY2024 | FY2023 | FY2022 |
|---|---|---|---|---|
| Revenue(high confidence) | $1,284M | $1,201M | $1,089M | $962M |
| Gross profit(high confidence) | $487M | $456M | $414M | $366M |
| Gross margin(high confidence) | 37.9% | 38.0% | 38.0% | 38.0% |
| EBITDA(medium confidence, derived value) | $218M | $203M | $184M | $163M |
| EBITDA margin(medium confidence, derived value) | 17.0% | 16.9% | 16.9% | 16.9% |
| EBIT(high confidence) | $176M | $164M | $148M | $131M |
| Net income(high confidence) | $124M | $115M | $103M | $91M |
| EPS (diluted)(high confidence) | $2.84 | $2.64 | $2.37 | $2.10 |
- High confidence
- Medium — derived value
- Low — verify against filing
Comparable companies
SIC-matched via SEC EDGAR · Market data via yfinance · EV in $M
| Company | Ticker | EV | EV/Rev | EV/EBITDA | EV/EBIT | P/E |
|---|---|---|---|---|---|---|
| Acme Corp (subject) | ACME | 2,180 | 1.7× | 10.0× | 12.5× | 21.2× |
| Meridian Software | MRSW | 3,420 | 2.8× | 13.6× | 16.2× | 26.4× |
| Crestline Systems | CRST | 1,870 | 1.5× | 9.4× | 11.8× | 19.1× |
| Harbinger Technologies | HRBG | 2,640 | 2.2× | 11.8× | 14.3× | 23.7× |
| Vantage Analytics | VTGA | 890 | 0.9× | 7.8× | 10.1× | 16.2× |
| Stratum Platforms | STRM | 3,150 | 3.1× | 15.2× | 18.7× | 29.0× |
| Median | 2,640 | 2.2× | 12.7× | 15.3× | 24.7× |
Capabilities
Built to hold up in a data room.
Confidence scoring
Every line item carries a high, medium, or low confidence marker. Medium means the value was derived from adjacent items. Low means it could not be reconciled — always verify against the source filing.
- Directly extracted from the filing
- Derived (e.g. EBITDA = EBIT + D&A)
- Could not reconcile — verify manually
- Machine-readable JSON confidence field
SEC EDGAR sources
Every number traces to an SEC filing, with an audit-ready citation trail. No scraped or estimated data. Peer discovery uses EDGAR's SIC-matched company index.
- EDGAR XBRL API for structured data
- SIC code peer matching — no manual selection
- Filing URL and line-item reference per row
- No estimated or model-interpolated values
Multi-pass verification
A second Claude pass checks the first extraction for sign errors, label mismatches, and unit inconsistencies. Disagreements between passes are flagged for human review.
- Independent extraction and cross-check
- Sign convention validation (debit / credit)
- Unit normalization to USD millions
- Disagreement flagging in the output CSV
CSV export
A two-section spreadsheet: subject company spread plus comps table, normalized in millions USD, with consistent column headers.
- IS, BS, and CF in a single structured sheet
- Comps table with EV, multiples, and ticker
- Consistent column headers across all outputs
- Ready for Excel, Google Sheets, or Python
Who uses it
For analysts who'd rather analyze.
Investment banking
Spread 10-Ks during due diligence, CIM prep, or fairness opinion work. Build the comps table before the Monday morning meeting — without spending your Sunday in CapIQ.
Private equity
Screen targets faster. Get IS / BS / CF plus trading comps on any public comparable in seconds. Spend your time on analysis, not data entry.
M&A advisory
Deliver structured financials and sector comps to clients without a CapIQ subscription. SEC EDGAR-sourced data holds up in a data room.
See your 10-K spread and comps in under a minute.
SEC EDGAR sourced · Confidence scored · CSV ready