Sells lenders usage-billed document automation (classification, data capture, tamper detection) plus cash-flow and income analytics, concentrated in small-business funding and mortgage, and is extending into mortgage underwriting conditions and GSE income checks.
What is Ocrolus, and which banks use it?
Ocrolus is a private New York company, founded in 2014, that classifies and extracts data from lending documents such as bank statements, pay stubs and tax forms, checks them for tampering and returns cash-flow and income analytics. Its last disclosed round was an $80 million Series C in 2021 at a valuation above $500 million. Among our 100 banks, Ocrolus names SoFi as a customer, in 2021 and 2023 releases. Pricing is not published; Ocrolus bills per book, form, document, application or funded loan.
Category
Founded
2014
Headquarters
New York, NY, US
Ownership
Private and venture-backed. Its $80 million Series C (September 2021) was led by Fin VC at a valuation 'north of $500 million', with Thomvest Ventures, Mubadala Capital, Oak HC/FT, FinTech Collective, QED Investors, Bullpen Capital and Cross River Bank among participants; Ocrolus now says it has raised $110 million in venture funding.[5][1]
Deployment
Cloud API and web dashboard, with loan-origination-system integrations (including ICE Encompass)
Pricing
Usage-based SaaS: each product is billed as a standalone SKU per Base Book, form, statement, page or document; mortgage tiers are billed per application, and tiers that include Analyze can be billed per funded loan · not published
Banks among the 100
1
- Ocrolus raised an $80 million Series C in September 2021, led by Fin VC, at a valuation it described as north of $500 million; its about page now says it has raised $110 million in venture funding.
- CEO Sam Bobley told TechCrunch in 2021 that revenue grew from $1 million to $20 million in annual recurring revenue between the first quarter of 2018 and the second quarter of 2021.
- Ocrolus's own releases name fintech and non-bank lenders as customers (Brex, Enova, LendingClub, PayPal, Plaid, Square, Kapitus) plus SoFi, the one bank among our 100 it names.
- Ocrolus says its platform analyzes roughly 750,000 credit applications a month and serves over 500 customers (vendor-reported, September 2026); its March 2026 release said over 400.
- Mortgage is described as its fastest-growing segment: more than 130 mortgage customers in September 2025, and nearly 90 mortgage lenders signed in the year to March 2026 (vendor-reported).
- Ocrolus's product documentation lists the billing unit for each product (per Base Book, form, statement, page, document, shared file, application or funded loan) but no prices.
- In 2025 and 2026 Ocrolus integrated with Fannie Mae's Income Calculator and Freddie Mac's AIM Check API, so extracted income data can be sent to the GSEs' own income assessments.
What does Ocrolus sell to banks?
| Product | What it does | Status |
|---|---|---|
| Classify[4][2] | Identifies and routes 2,100+ document types (bank statements, pay stubs, tax returns, W-2s, mortgage applications, IDs) before extraction; billed per Base Book or form. | Generally available |
| Capture (Bank Statement Analysis, Income, Tax Forms, Identification)[4] | Field-level extraction from bank statements, pay stubs, W-2s, 1099s, 1040s and identity documents, which Ocrolus says runs at 99%+ accuracy; billed per statement, page, Base Book or form. | Generally available |
| Detect[3][2] | Checks documents for manipulation, tampering and metadata anomalies such as abnormal fonts, invalid dates and irregular formats; billed per document. | Generally available |
| Analyze (cash flow and income)[2][10][15] | Cash-flow analytics for small-business funding and income calculation for mortgage, with integrations to Fannie Mae's Income Calculator (2025) and Freddie Mac's AIM Check API (2026). | Generally available |
| Inspect[8] | Mortgage data validation that flags discrepancies between borrower documents and application data, integrated with ICE's Encompass loan origination system; introduced October 2024. | Generally available |
| Automated conditioning[12] | Generates, tracks and resolves mortgage underwriting conditions in one workspace with Encompass sync; announced as generally available from April 1, 2026. | Generally available |
| Encore[11][3] | A double opt-in platform for brokers and small-business funders to share borrower files and standardized cash-flow data; launched October 2025, billed per shared file. | Generally available |
How much does Ocrolus cost?
How it is charged: Usage-based SaaS: each product is billed as a standalone SKU per Base Book, form, statement, page or document; mortgage tiers are billed per application, and tiers that include Analyze can be billed per funded loan.
Ocrolus does not publish a price list.
Pricing is not published; Ocrolus's product SKU reference documents the billing unit for each product but lists no prices, and contracts are set in a master services agreement.
Which banks use Ocrolus?
| Bank | What | Status |
|---|---|---|
| SoFi Technologies | Named client (2021, 2023): Ocrolus said in 2021 that it 'powers document workflows' for SoFi and named SoFi again in a May 2023 release; no later confirmation found.[5][7][6] | Selected |
Among the 100 largest US banks on this site, from sources that name the bank: the bank, Ocrolus or the press. Status is the furthest stage a source confirms.
What is Ocrolus platform (Classify, Capture, Detect, Analyze)?
An API and dashboard that classifies uploaded documents, captures their data fields, checks qualifying documents for tampering and analyzes borrower cash flow and income; documents the models cannot resolve are routed to Ocrolus's own verifiers for human review.[2]
What has Ocrolus done, and when?
- Sep 10, 2026
Integrates Freddie Mac's AIM Check API[15]
Lenders using Ocrolus Analyze can send wage-earner income from W-2s and pay stubs to Freddie Mac's asset and income modeler for an early income assessment before a full Loan Product Advisor submission.
- Jul 15, 2026
Documents product SKUs and billing units[3]
Ocrolus's documentation lists every product SKU with its unit of measure (reference last updated July 15, 2026), stating that each SKU name matches the billing identifier in a customer's master services agreement.
- Mar 17, 2026
Automated mortgage conditioning goes generally available[12]
Ocrolus announced general availability from April 1, 2026 of automated condition generation, tracking and resolution with Encompass sync, built using feedback from mortgage customers including American Federal Mortgage and Zillow.
- Oct 28, 2025
Launches Encore for small-business funding[11]
Encore lets brokers and funders share borrower files and standardized cash-flow metrics; it opened with early adopters, with broader availability planned for January 2026.
- Sep 15, 2025
Integrates Fannie Mae's Income Calculator[10]
Extracted income data for self-employed and rental-income borrowers is sent to Fannie Mae's Income Calculator, whose results can make loans eligible for representation-and-warranty relief on the income calculation.
- Jun 25, 2025
Splits into Mortgage and SMB business units[9]
Nadia Aziz joined as General Manager of Mortgage and David Snitkof was promoted to General Manager of SMB; Ocrolus called mortgage its fastest-growing segment and SMB funding its largest line of business.
- Oct 28, 2024
Introduces Inspect for mortgage data validation[8]
Inspect automates the comparison of borrower documents with application data and integrates with ICE's Encompass; Ocrolus said it then had over 100 mortgage lending customers.
- May 8, 2023
Adds OpenAI GPT embeddings[7]
Ocrolus integrated OpenAI GPT embeddings alongside Amazon, Google and its own models to process semi-structured documents without human involvement.
- Sep 23, 2021
- Jan 1, 2014
What do banks use Ocrolus for?
| Use | What it does | Status | Use case |
|---|---|---|---|
| Small-business cash-flow underwriting[11][9] | Bank-statement extraction and cash-flow analytics that funders feed into credit scoring and underwriting of small-business loans. | In production | Credit scoring & underwriting |
| Mortgage income calculation and validation[8][10][15] | Income calculation, document-to-application checks (Inspect) and GSE income assessments through Fannie Mae and Freddie Mac integrations. | In production | Credit scoring & underwriting |
| Document tampering detection[3][2][14] | Detect flags edited or anomalous bank statements, pay stubs and other documents before underwriting. | In production | Fraud detection |
| Automated underwriting conditions[12] | Generates mortgage underwriting conditions from asset, income and credit data and matches incoming documents to them; underwriters review, edit and import the conditions. | Rolling out | Generative & agentic AI |
What has Ocrolus disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Credit applications analyzed per month (vendor-reported) | Roughly 750,000[15] | Sep 10, 2026 |
| Customers (vendor-reported) | Over 500 (the March 2026 release said over 400)[15][12] | Sep 10, 2026 |
| Mortgage customers (vendor-reported) | More than 130[10] | Sep 15, 2025 |
| Series C funding and valuation | $80 million at a valuation north of $500 million; total raised over $100 million at the time[5][6] | Sep 23, 2021 |
| Annual recurring revenue (per the CEO, 2021) | $20 million, up from $1 million in Q1 2018[6] | Jun 30, 2021 |
| Compute cost per document (vendor-reported comparison) | About 3 cents on Ocrolus's specialized models versus about 30 cents on a managed large language model provider[13] | Jul 21, 2026 |
| Share of small-business applications with a suspicious document (vendor-reported) | Roughly 1 in 8, across 1.46 million applications over a recent 90-day period[14] | Sep 3, 2026 |
| Small-business credit applications per month (vendor-reported) | Close to 500,000[9] | Jun 25, 2025 |
“Our platform helps lenders automate underwriting and intelligently leverage cash flow and income data for credit scoring.”[5]
Sam Bobley · Co-founder and CEO, Ocrolus · Sep 23, 2021
“With this integration, that same data can now power an early income assessment from AIM inside lenders' existing workflows.”[15]
Nadia Aziz · General Manager of Mortgage, Ocrolus · Sep 10, 2026
Who runs Ocrolus?
Which rules apply to a bank using Ocrolus?
| Authority | Why it matters here | Documents |
|---|---|---|
| CFPB | Where a bank uses Ocrolus's extracted income and cash-flow figures to approve or decline credit, ECOA/Regulation B still requires the bank to give specific, accurate adverse-action reasons, including when the decision rests on complex models (CFPB Circular 2022-03). | ECOA / Regulation B adverse action (15 U.S.C. 1691(d); 12 CFR 1002.9) CFPB Circular 2022-03 |
| OCC | Ocrolus processes borrowers' bank statements, pay stubs and tax forms and routes some documents to its own human verifiers, so a national bank using it must manage it as a third party under OCC Bulletin 2023-17, and bring any analytics that feed credit decisions under the revised model-risk guidance (OCC Bulletin 2026-13). | OCC Bulletin 2023-17 OCC Bulletin 2026-13 |
| Federal Reserve | For a Fed-supervised bank, Ocrolus's income calculations and cash-flow metrics are vendor-supplied quantitative outputs used in underwriting; SR 26-2 expects validation and ongoing monitoring of such vendor models, and SR 23-4 covers the third-party relationship. | SR 26-2 SR 23-4 |
| FDIC | State non-member banks outsourcing document review and fraud checks to Ocrolus fall under the interagency third-party guidance issued by the FDIC as FIL-29-2023, covering due diligence, contract terms on data use and ongoing monitoring. | FDIC FIL-29-2023 |
Linked authorities have a page in the AI Regulation Tracker; the others are named for completeness.
What should a bank note before buying Ocrolus?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice or a recommendation.
01
Pricing is not published, but billing units are
Ocrolus's documentation names the unit each product is billed on (per Base Book, form, page, document, application or funded loan) without prices; any comparison of cost needs a quote against the buyer's own document mix.
02
The one named bank customer rests on older releases
SoFi is named as a customer in Ocrolus's 2021 Series C release and its 2023 OpenAI release; no 2024-2026 source reviewed here repeats it, and most customers Ocrolus names are fintechs, non-bank lenders and mortgage companies.
03
Customer counts differ between releases
Ocrolus's boilerplate said over 500 customers in October 2024, over 400 in June 2025 and March 2026, and over 500 again in September 2026; the figures are vendor-reported and not reconciled.
04
Accuracy and cost claims are Ocrolus's own
The 99%+ accuracy figure and the claim that its specialized models cost about 3 cents per document in compute versus about 30 cents on a managed LLM come from Ocrolus's own documentation and blog, not an independent test.
05
Humans still review some documents
Ocrolus's documentation says documents its models are unsure of are routed to Ocrolus verifiers for human review, which matters for a bank's data-handling and third-party due diligence.
What are the alternatives to Ocrolus?
Blend
Origination software for mortgages, home equity, consumer loans and deposit accounts; its Autopilot AI agent reviews borrower documents.
Snapdocs
Digital mortgage closing platform (eClosing, eVault, notary network) with AI quality control of closing and funding documents.
Abrigo
Lending, loan review, CECL and AML software for US community and regional banks, with generative and agentic AI assistants added since 2025.
Covecta
Agentic AI for banks and lenders: pre-built agents that read documents, draft credit memos and run loan and deposit operations checks.
How much does Ocrolus cost?
Ocrolus does not publish prices. Its product SKU reference says each product is billed and contracted separately, per Base Book, form, statement, page or document for core products, per shared file for Encore, and per application (or per funded loan for tiers including Analyze) for mortgage. The actual rates are set in each customer's master services agreement.
Which banks use Ocrolus?
Among the 100 largest US banks we track, Ocrolus names SoFi as a customer, in its September 2021 Series C release and its May 2023 OpenAI release. Most customers Ocrolus names publicly are fintech and non-bank lenders such as Brex, Enova, PayPal, Square and Kapitus, plus mortgage lenders such as CrossCountry Mortgage and Better, and smaller banks such as HomeTrust Bank.
Who owns Ocrolus?
Ocrolus is private and venture-backed. Fin VC led its $80 million Series C in September 2021 at a valuation above $500 million, with Thomvest Ventures, Mubadala Capital, Oak HC/FT, FinTech Collective, QED Investors, Bullpen Capital and Cross River Bank participating. Ocrolus says it has raised $110 million in venture funding in total; no later priced round was found.
Does Ocrolus detect fraudulent documents?
Yes, through its Detect product, which Ocrolus says checks qualifying documents for tampering, incomplete data sets, invalid dates and amounts, abnormal fonts and irregular formats. Detect is billed per document. Ocrolus reports that roughly one in eight small-business lending applications contains a suspicious document, which is its own figure.
Is Ocrolus a bank or regulated lender?
No. Ocrolus is a software and data vendor that processes documents for lenders. The banks and lenders that use it make the credit decisions and remain responsible for fair-lending, adverse-action and privacy obligations; for a bank, Ocrolus is a third party to be managed under the interagency third-party risk guidance.
Ocrolus vs Blend: what is the difference?
They sit at different points of the loan process. Blend sells the consumer-facing application and origination workflow (mortgage, home equity, deposits) on per-transaction fees. Ocrolus processes the documents and data behind underwriting, such as bank statements, pay stubs and tax forms, and returns extracted data, tamper checks and cash-flow or income analytics. A lender can use both.
- About Ocrolus | Our Vertical AI Journey (undated page, retrieved 2026-09-24)
- How Ocrolus Works | Intelligent Document Processing Explained (undated page, retrieved 2026-09-24)
- Ocrolus product SKUs: Overview
- Core SKUs | Ocrolus product SKUs
- Ocrolus Raises $80M In Series C Funding To Scale Its Financial Services Focused Document Automation Solution
- Ocrolus lands $80M at a $500M+ valuation to automate document processing for fintechs and banks
- Ocrolus Adds OpenAI GPT Embeddings for Deeper Automation in Financial Document Analysis
- Ocrolus introduces Inspect for automated mortgage data validation
- Ocrolus appoints Nadia Aziz to lead Mortgage business, promotes David Snitkof to oversee SMB
- Ocrolus engages with Fannie Mae to streamline mortgage income calculations
- Ocrolus launches Encore: A first-of-its-kind, trusted cash flow data sharing platform for small business funding
- Ocrolus accelerates automated conditioning for mortgage lenders with full lifecycle management
- What AI-powered loan origination actually costs in 2026
- Automated bank statement analysis: how Banclease cut review time by 90%
- Ocrolus integrates with Freddie Mac's AIM Check API to offer an automated income assessment solution
When a bank signs with Ocrolus, you'll read it here first.
when a bank picks an AI vendor, or a vendor ships for banks, it is in the next morning's brief · six sourced stories · 7 am ET · free
plus every tracker, bank and agent page update, the morning after · leave any morning