Every value from every filing, with the sentence that proves it
Lucy licenses Lucy RAG and 17 structured datasets built from SEC filings to AI companies, research platforms, data vendors, and funds. Each license is shaped around how you use the data. Tell us your use case and we'll send a written proposal.
NVIDIA · largest direct customer
22%
of total revenue · threshold 10%
“For fiscal year 2026, sales to one direct customer represented 22% of total revenue and sales to another direct customer represented 14% of total revenue…”
- accession
- 0001045810-26-000021
- filed
- 2026-02-25 · EDGAR
- 17
- structured datasets
- 2016–today
- of SEC filing history
- 94.7%
- Recall@10 on FinanceBench
- +0.160
- answer accuracy from retrieval alone
Two ways to use Lucy, one license
Both are built from the same filings, with the same evidence on every value. Most teams license both.
Questions in plain language
Lucy RAG
Send a question and get back the passages, tables, and XBRL facts that answer it, each with its accession number and section. Your model writes the answer.
See use casesData, field by field
Structured datasets
17 datasets, from properties to insider transactions, with the source text behind every value. Ready for tables, screens, and models.
Browse the Data CatalogHow licensing works
Most teams take four to eight weeks from evaluation to a signed agreement.
Step 1 · Evaluate
Free · 30 days
Up to 100 companies you choose, with full history, every dataset, and Lucy RAG. You get a data dictionary, an accuracy report, and a kickoff with a solutions engineer.
Step 2 · Pilot
90 days · credited
Production use in one product or workflow, with full coverage. The pilot fee is credited in full toward your first-year license.
Step 3 · License
Annual or multi-year
An agreement with the use rights, coverage, history, and delivery your use case needs.
Choose a license for who sees your data
Every license includes the evidence on every value, the REST API, and the MCP server. The license type decides who may see the data and what you may do with it.
Internal research
For your own teams
Analysts, internal models, and internal tools, including AI assistants your employees use.
Funds, research, and strategy teams
Product display
For your customers
Values and AI answers shown to your users inside your product, with attribution.
Research platforms, AI assistants for investors
Redistribution
For resale
The data itself delivered to others as feeds, files, or exports, or inside products you sell.
Data vendors, terminals, marketplaces
Add-on · AI training
For your models
Training or fine-tuning your own models on Lucy data. Added to any license.
Model builders, AI product teams
What each license allows
Use rights are spelled out in your order form and our Terms of Service.
Swipe to compare all three licenses, or focus the table and use arrow keys.
| Use rights | Internal research | Product display | Redistribution |
|---|---|---|---|
| Analysts, internal models, and tools use the data | Included | Included | Included |
| Your AI assistant answers employees with Lucy evidence | Included | Included | Included |
| Show values and AI answers to your customers | Not included | Included | Included |
| Deliver the data itself to others as feeds, files, or exports | Not included | Not included | Included |
| Attribution ("Source: Lucy") next to displayed values | Not needed | Required | Required |
| Train or fine-tune your own models | Add-on | Add-on | Add-on |
What shapes your price
We don't publish an enterprise price list, because a research team and a terminal serving 250,000 users need very different agreements. These are the factors we price on. After a scoping call, you get a written proposal within one business day.
Discuss your use case- License type
- Internal research, product display, or redistribution
- Products
- Lucy RAG, structured datasets, or both
- Audience size
- How many end users can see the data
- Retrieval volume
- Annual Lucy RAG requests, for retrieval licenses
- History and delivery
- Recent years or full history, bulk delivery, uptime SLA
- Term
- Multi-year agreements lock in pricing
Who licenses Lucy
The license you need depends on who sees the data. These are the four most common setups.
Product display
AI assistants for investors
Ground answers in filing evidence your model can cite, over the REST API or MCP.
Display or redistribution
Research platforms
Add structured, source-linked fields to company pages, search, and screeners.
Redistribution
Data vendors and terminals
Add evidence-linked US disclosures to the feeds your clients already use.
Internal research
Funds and research teams
Screen, backtest, and research company disclosures, with history from 2016.
17 datasets from SEC filings
Built from 10-K and 10-Q reports, DEF 14A proxy statements, and Form 4 insider filings. Every value ships with the filing it came from, the section it was found in, and the sentence behind it.
PROP · PROD · HC · IND
Company Profile
Properties and facilities, product portfolio, workforce, industry classification
FIN · REV
Financials
Financial statements, income statement flow, income statement commentary
CAP
Capital Structure & Allocation
Acquisitions, debt instruments, share repurchases
RISK
Risk & Legal
Legal proceedings and customer concentration, including explicit non-disclosures
INSD · OWN · MGMT
Ownership & Governance
Insider transactions, beneficial ownership, officers and directors
TL · SUM
Filing Activity
Filing timeline and filing summaries
Also available
Beyond public companies
- Funds & ETFs · coming soon
- International coverage · available on request
Same schema, evidence model, and delivery as SEC data. Tell us what you need on the form below.
Each record carries its source
A real record from NVIDIA's fiscal 2026 annual report, from the Customer concentration dataset. Your analysts can check any value against the filing, and your model can cite it.
- evidence.quote
- The verbatim sentence from the filing
- evidence.reasoning
- Why the value was extracted this way
- accession_number
- The exact SEC filing
- form · filing_date
- Form type and when it was filed
- report_date
- The period the value describes
- disclosure_status
- Disclosed, explicitly none, or not disclosed
// SEC · customer_concentration · NVIDIA 10-K, fiscal 2026
{
"metadata": {
"form": "10-K",
"accession_number": "0001045810-26-000021",
"filing_date": "2026-02-25",
"report_date": "2026-01-25"
},
"values": {
"disclosure_status": { "value": "disclosed" },
"disclosure_threshold": { "value": 10.0, "unit": "percent" },
"concentrations": [{
"category": "revenue", "scope": "single",
"value": { "amount": 22.0, "unit": "percent" },
"evidence": {
"quote": "For fiscal year 2026, sales to one direct customer represented 22% of total revenue…",
"reasoning": "One direct customer represented 22% of total revenue in fiscal 2026."
}
}]
}
}Coverage and delivery
What a license covers today, and what is coming next.
- Structured datasets
- 17 datasets from 10-K, 10-Q, 20-F, 40-F, DEF 14A, and Form 4 filings
- Lucy RAG retrieval
- 10-K, 10-Q, 8-K, DEF 14A, Forms 3, 4, and 5, 13F, 13D/G, and more, as retrieval-ready passages, tables, and XBRL facts
- Universe
- Companies filing with the SEC, including foreign private issuers
- History
- 2016 to today
- Updates
- New EDGAR filings processed in near real time
- Evidence
- Accession number, section, source quote, and reasoning on every value, or the XBRL concept for financial statement figures
- Delivery
- REST API and MCP server. Bulk delivery available on request.
- Funds & ETFs
- Coming soon
- International
- Available on request
API plans & usage limits
API plans set your request and token limits. Enterprise licenses define data use rights, audience, coverage, and delivery. API access alone does not grant product display, redistribution, or AI training rights; those are agreed in your license.
Ready for procurement
What your legal, security, and data teams will ask for, prepared before the first call.
Master agreement and order form
Standard terms, with rights spelled out per license
Data Processing Addendum
For customers who send us personal data
Security review
SOC 2 Type II audit in progress, plus a completed security questionnaire
Data dictionary
Every field, with types, definitions, and examples
Accuracy report
Field-level results on a hand-checked set of filings
Uptime SLA
Available on product display and redistribution licenses
Questions enterprise buyers ask
Licenses are priced on how you use the data, the products you need, and how many people will see it. A research team and a terminal with 250,000 users need very different agreements. After a 30-minute scoping call, we send a written proposal within one business day.
Yes. Start with a free 30-day evaluation on up to 100 companies, then a 90-day paid pilot. The pilot fee is credited in full toward your first-year license.
By license type and your expected annual retrieval volume. Evaluations include enough requests to test Lucy RAG on your own questions.
Yes, with a product display license. Answers must show "Source: Lucy" near the value. Training or fine-tuning your own models on Lucy data needs the AI training add-on.
Our SOC 2 Type II audit is in progress. During an evaluation, we share our security documentation and a completed security questionnaire.
Lucy covers every company that files with the SEC, including foreign private issuers that file 20-F and 40-F reports. Coverage of other markets is available on request.
Your right to use Lucy data ends with the term, as set out in our Terms of Service. We agree any wind-down period in your order form.
Licenses renew as set out in your order form. We give at least 30 days' notice of any price change, and multi-year terms lock in pricing.
Talk to sales
Get a proposal for your use case
Tell us how you'll use the data. We'll reply within one business day to set up a scoping call.
- A 30-minute call to confirm your use case, products, and license type
- A written proposal within one business day
- A free 30-day evaluation on companies you choose
- A pilot or annual license, whichever fits
Prefer email?
Sales and support
help@lucydata.ai