Every subtotal
ties out, or
it’s a flag.
Filing Studio reads the toolkit’s build artifacts and turns every disagreement between a filer’s disclosed number and the rendered model into a clickable card. Pick a vintage, fold a renamed tag, or escalate — git remembers your judgment.
- AAPLApple Inc.10-K · FY24↓ .xlsx
- MSFTMicrosoft Corp.10-K · FY24↓ .xlsx
- NVDANVIDIA Corp.10-K · FY25↓ .xlsx
- GOOGLAlphabet Inc.10-K · FY24↓ .xlsx
- TSLATesla Inc.10-K · FY24↓ .xlsx
Built for analysts
who show their work.
Six primitives that turn an SEC filing into a peer-reviewed thesis: a verifiable baseline, your assumptions, the community’s verdict, and a track record that follows you.
SEC Baselines
Auto-generated three-statement models from EDGAR XBRL filings. Every cell ties back to a tagged number in the filing.
Format verified
Uploads scored against the SEC baseline by line-item conformance. Green check = same chart of accounts.
- Income statement✓ 24 / 24
- Balance sheet✓ 31 / 31
- Cash flow✓ 19 / 19
Fork & layer
Take a baseline. Bend assumptions. Post the call with the .xlsx attached.
Track record
Locked calls. Bayesian-shrunk hit rate. The leaderboard rewards alpha, not volume.
On-demand baselines
Type a ticker. The XBRL pipeline ships a fresh three-statement model in under two minutes.
- •EDGAR fetch · 10-K FY25
- Parse XBRL · 1,284 tags
- Build statements · 74 line items
- Upload .xlsx · 2.4 MB
Community ranked
Reddit-mechanics for thesis discovery. Upvote what holds up. Download what helps.
- ▲412risingMETA▲ Bull
Reels CPM inflection by FY26
- ▲287—TSLA▼ Bear
Mix-shift to lower ASP vehicles
- ▲198—AVGO▲ Bull
Custom silicon TAM expansion
Three steps from a filing
to a track record.
Start at the truth.
8 SEC baseline models, auto-built from XBRL filings. Search any ticker, filter by sector, see how many community models already exist for it.
- Income / balance / cash flow — pre-wired
- Every cell tied to a tagged XBRL fact
- Missing a company? Type the ticker, get it in ~2 min
Layer on your thesis.
Bend any baseline to your view. Pick a stance, post the assumptions you changed, attach the .xlsx. The community sees exactly what's the SEC's and what's yours.
- Up to 6 labeled assumption chips with direction
- Format-verified against the original baseline
- Fair value + thesis locked at post time
Build a track record.
Upvotes surface what holds up. Resolved calls feed your shrunk hit-rate. The leaderboard rewards alpha across calls — not loud, not lucky.
- Bayesian-shrunk hit rate (the ranking metric)
- Locked calls are immutable once posted
- Followers, downloads, comments all roll up
Data-center capex cycle has another two years to run.
Capex /R ratio is the one I keep coming back to. Solid take.
How are you handling networking attach in FY26?
The best calls
rise to the top.
Locked at post time. Scored against the benchmark on resolve. Bayesian-shrunk so loud doesn’t beat right — and a streak of 4 / 5 doesn’t outrank 18 / 28. Track record is the only rep that compounds.
Dying giant? Or waking giant.
The record
is open.
Sign up, fork a baseline, post a thesis with assumptions tied to the filings. Your calls lock at post time. Your record runs forever.