THE RECORD · WITH FACTS

Checked against what runs.

AI-native companies and operators, their claims checked against reality with the receipts attached. 74 on the record, 310 of 586 claims verified.

11x.ai

5 of 8 claims verified

3 of 8 key claims verified, 3 contradicted by independent reporting, 1 contradicted with company dispute, 1 gap on current 2026 figures

View record

Abridge

4 of 10 claims verified

4 of 10 claims verified, 2 plausible, 4 thin, 0 contradicted. No confirmed 2026 revenue figure and no official headcount disclosure.

View record

Ambience Healthcare

4 of 7 claims verified

4 of 7 key claims verified, 1 plausible, 2 thin

View record

Artisan AI

5 of 10 claims verified

5 of 9 claims verified, 2 thin, 1 plausible, 1 contradicted

View record

AssemblyAI

3 of 6 claims verified

3 of 6 key claims verified, 2 thin (self-reported or single-source), 1 unresolved (ARR figure could not be traced to primary source)

View record

Assort Health

3 of 7 claims verified

3 of 7 key claims verified, 1 plausible, 3 thin

View record

Base44

7 of 8 claims verified

7 of 8 key claims verified, 1 contradicted (solo-run exit framing). $38M of the $90M earnout already paid as of Q1 2026.

View record

Baseten

6 of 8 claims verified

5 of 7 key claims verified, 1 plausible (single-source), 1 thin (headcount)

View record

Black Forest Labs

6 of 10 claims verified

6 of 10 claims verified, 2 plausible, 2 thin/forecast

View record

Bland AI

3 of 7 claims verified

3 of 7 key claims verified, 3 thin, 1 confirmed controversy

View record

Bolt.new

3 of 7 claims verified

3 of 7 key claims verified (founding, Series B funding, Microsoft partnership); 3 plausible (revenue growth, profitability, lean/agent-heavy headcount, all founder self-reported); 1 thin (75% Fortune 500 adoption)

View record

Cal AI

4 of 7 claims verified

4 of 7 key claims verified, 1 plausible (self-reported revenue), 2 contradicted (headcount conflict, accuracy claim vs independent testing)

View record

Captions AI (Mirage)

5 of 10 claims verified

5 of 10 claims verified via primary or reputable sources, 3 plausible (company-reported, not independently audited), 2 thin (model-estimated revenue and conflicting headcount figures)

View record

Chai AI

3 of 8 claims verified

3 of 8 key claims verified, 2 plausible, 2 thin, 1 contradicted

View record

Character.AI

7 of 10 claims verified

7 of 10 claims verified via fetched sources; 3 (monthly active users, revenue, headcount) are thin, conflicting estimates from secondary aggregators since the company is private and does not publish audited metrics.

View record

Clay

3 of 9 claims verified

3 of 9 claims verified, 3 plausible, 3 thin. Core revenue and funding milestones confirmed via company blog plus independent press; headcount and internal efficiency metrics are self-reported or contradictory across sources.

View record

CodeRabbit

4 of 8 claims verified

3 of 8 key claims verified, 2 plausible, 2 thin, 1 contradicted (third-party data source, not CodeRabbit itself)

View record

Cognition (Devin)

5 of 7 claims verified

5 of 7 key claims verified; 1 thin (90% internal code, self-reported only); 1 plausible (performance benchmarks, self-reported with partial customer corroboration)

View record

Crescendo AI

2 of 6 claims verified

2 of 6 key claims verified, 1 plausible, 1 thin, 2 contradicted by headcount and positioning reality

View record

Cresta

5 of 8 claims verified

6 of 8 key claims verified, 1 plausible/stale, 1 thin

View record

CrewAI

3 of 9 claims verified

3 of 8 key claims verified via primary sources, 5 thin (self-reported/aggregator-only, no independent confirmation)

View record

Cursor (Anysphere)

4 of 7 claims verified

5 of 7 key claims verified; 1 plausible; 1 thin. SpaceX acquisition confirmed but not yet closed.

View record

Decagon

7 of 10 claims verified

6 of 8 key claims verified, 2 thin (current ARR estimate, headcount)

View record

Deepgram

6 of 10 claims verified

6 of 9 key claims verified, 3 thin (self-reported benchmark, founding date, and headcount not confirmed via a primary source)

View record

ElevenLabs

7 of 8 claims verified

7 of 8 key claims verified, 1 thin. Primary sources: company blog, TechCrunch, CNBC, U.S. Senate JEC. Minor ARR discrepancy flagged.

View record

EliseAI

3 of 7 claims verified

3 of 7 key claims verified, 1 plausible, 3 thin

View record

Emergent

3 of 6 claims verified

4 of 6 key claims verified, 1 contradicted (headcount conflict), 1 contradicted (reliability claims vs. user reports)

View record

EvenUp

2 of 7 claims verified

2 of 7 key claims verified, 1 plausible, 4 thin

View record

Exa

2 of 6 claims verified

2 of 6 key claims verified, 2 plausible, 2 thin

View record

Factory

5 of 9 claims verified

5 of 8 key claims verified, 2 thin, 1 contradicted (headcount data conflict)

View record

Fal.ai

4 of 6 claims verified

4 of 6 key claims verified, 1 plausible (pending), 1 thin (headcount)

View record

Fireflies.ai

3 of 8 claims verified

3 of 8 key claims verified, 3 plausible, 1 thin, 1 contradicted by pending litigation

View record

Fyxer AI

5 of 11 claims verified

5 of 11 key claims verified, 3 plausible, 3 thin; 2 forward targets correctly labeled as forecasts

View record

Gamma

5 of 6 claims verified

5 of 6 key claims verified; 1 thin (2026 headcount after Series B hiring spree)

View record

Genspark

4 of 7 claims verified

4 of 7 key claims verified, 1 plausible, 1 thin, 1 contradicted (headcount data conflicts with stated 100-employee cap)

View record

Glean

5 of 8 claims verified

5 of 8 key claims verified, 2 thin, 1 plausible; notably Glean's own CEO publicly rejects the lean/near-zero-employee AI-native model, projecting headcount growth to 5,000-10,000 within five years

View record

Gradient Labs

4 of 11 claims verified

4 of 11 key claims verified, 2 plausible, 5 thin; headcount undisclosed by the company (gap)

View record

Granola

6 of 8 claims verified

6 of 8 key claims verified, 1 plausible (founding date), 1 thin (revenue growth percentage); the founder-stated team size of roughly 70 to 80 employees contradicts a third-party aggregator's estimate of 116 to 127.

View record

Harvey

3 of 7 claims verified

3 of 7 key claims verified, 2 plausible, 2 thin

View record

Hebbia

5 of 9 claims verified

5 of 9 key claims verified, 4 thin. Core Series B, revenue-at-raise, asset-manager penetration, and FlashDocs acquisition are confirmed; current headcount, later ARR, usage-scale numbers, and OpenAI-volume claim rest on single or first-party sources only.

View record

HeyGen

2 of 6 claims verified

2 of 6 key claims verified, 1 plausible, 2 thin, 1 contradicted

View record

Higgsfield AI

7 of 11 claims verified

6 of 10 key claims verified, 2 forecast/pending, 1 thin (headcount), 1 verified controversy

View record

Hippocratic AI

3 of 6 claims verified

3 of 6 key claims verified, 2 thin, 1 plausible; headcount and revenue figures are gaps

View record

Instantly.ai

0 of 7 claims verified

0 of 7 tracked claims fully verified; 4 plausible, 2 thin, 1 contradicted (self-reported customer count vs. independent paying-customer estimates)

View record

Krea AI

3 of 7 claims verified

3 of 7 key claims verified (founding/founders, funding/valuation, Krea 2 leaderboard rank); 4 thin (ARR, user count, lean-team claim, enterprise customer logos), all resting on self-reported or single-source data

View record

Legora

4 of 6 claims verified

4 of 6 key claims verified, 1 plausible, 1 thin (investor-sourced superlative)

View record

Lindy AI

2 of 8 claims verified

2 of 8 claims verified, 3 plausible, 2 thin, 1 contradicted by the company's own inconsistent marketing figures

View record

Lovable

7 of 8 claims verified

5 of 6 key claims verified against primary sources; 1 ARR figure is analyst estimate; notable security CVE and VAT controversy documented

View record

Luma AI

2 of 9 claims verified

2 of 9 key claims verified (Series C round, HUMAIN/Riyadh partnership), 2 plausible (founding story, total funding/unicorn status), 5 thin (revenue, users, headcount, Ray3 technical/benchmark claims, Adobe integration). No contradictions found, but most quantitative claims rest on self-reported or aggregator-only sourcing.

View record

Manus

3 of 7 claims verified

3 of 7 claims verified, 1 contradicted, 3 thin

View record

Mercor

4 of 9 claims verified

4 of 9 claims verified, 2 plausible, 1 thin, 2 contradicted. Growth and funding numbers check out; the AI-native/lean-headcount framing does not.

View record

Midjourney

5 of 9 claims verified

Self-funded status and early profitability verified via primary sourcing. Revenue figures are analyst estimates, not company disclosures. Headcount figures conflict significantly across sources. $10B valuation is a derived estimate, not a disclosed figure. 3 of 8 key claims fully verified, 3 plausible, 1 thin, 1 contradicted.

View record

n8n

5 of 9 claims verified

5 of 8 key claims verified, 2 thin, 1 forecast (not established fact)

View record

OpenArt AI

3 of 5 claims verified

3 of 5 key claims verified, 1 plausible (self-reported ARR not independently audited), 1 contradicted (lean-team/revenue-per-employee framing)

View record

Opus Clip (OpusClip)

1 of 8 claims verified

1 of 8 key claims fully verified (SoftBank funding round), 3 plausible (cumulative funding, usage scale, founding details), 4 thin (revenue/ARR figures, headcount, the '$1M ARR in 14 days' origin story, and the 'world's first' ClipAnything superlative)

View record

Otter.ai

3 of 8 claims verified

3 of 8 claims verified, 1 plausible, 3 thin (self-reported scale figures), 1 contradicted (a widely-repeated '$1B unicorn' claim not supported by primary funding records)

View record

Parloa

5 of 7 claims verified

5 of 7 key claims verified, 2 plausible (self-reported metrics)

View record

Perplexity

5 of 7 claims verified

5 of 6 key claims verified, 1 thin (headcount); 2 confirmed legal controversies

View record

Photoroom

4 of 7 claims verified

4 of 7 key claims verified, 2 thin, 1 contradicted (headcount claims conflict with company's own lean-team positioning)

View record

Replit

4 of 10 claims verified

4 of 9 claims verified, 2 plausible, 2 thin, 1 contradicted; the $1B 2026 revenue figure is a company forecast, not an established fact

View record

Retell AI

3 of 5 claims verified

3 of 5 key claims verified, 1 thin, 1 plausible

View record

Rilla

3 of 7 claims verified

2 of 7 key claims verified, 5 thin (single-source estimates)

View record

Rogo

5 of 8 claims verified

5 of 8 key claims verified, 1 thin, 2 plausible, founding date unresolved (gap)

View record

Runway

8 of 10 claims verified

7 of 10 key claims verified, 1 plausible (revenue), 1 thin (headcount), 1 verified-with-caveat (benchmark claim since superseded)

View record

Sierra

5 of 7 claims verified

4 of 6 key claims verified via primary sources, 2 thin (May 2026 ARR figure and headcount unconfirmed)

View record

Speak

6 of 9 claims verified

6 of 9 claims verified, 1 contradicted (headcount), 1 thin (Korea #1 ranking), 1 plausible (derived revenue-per-employee ratio)

View record

Suki AI

4 of 10 claims verified

4 of 10 key claims verified, 5 plausible, 1 thin (self-reported revenue growth); no lawsuits or contradictions found

View record

Suno

3 of 6 claims verified

3 of 6 key claims verified with primary sources, 1 plausible (self-reported metric), 1 thin (conflicting headcount data), 1 contradicted (a widely repeated $500M figure misstates the WMG settlement)

View record