Document and regulation intelligence
We turn rules, contracts and filings into records your systems can check. Each field links to the passage it came from, and a person reviews every record.
What we work with
Regulatory intelligence and contract data extraction with LLMs: obligations, triggers, thresholds, exceptions and deadlines read from the text into records, each linked to the phrase it came from.
| Document type | What we extract | What it becomes |
|---|---|---|
| Legal decisions and statutes | Holdings, outcomes, dates, cited authorities and the passages that support them | Research corpora, citation graphs and retrieval for legal research agents |
| Regulation | Obligations, triggers, thresholds, exceptions and effective dates | Obligation registers and rule models |
| Exchange and venue rulebooks | Order types, priority and matching rules, auctions, price bands and halts | Machine-checkable models of order handling |
| Technical specifications | Parameters, limits, interfaces and conformance requirements | Requirement tables and conformance checks |
| Contracts and filings | Parties, defined terms, obligations, conditions and deadlines | Structured records and obligation and deadline lists |
Scanned pages are read with OCR, and each field still links to its page and passage.
Contracts and filings
Parties, defined terms, obligations, conditions and deadlines, read into one record per contract or filing. Each field links to its passage and page, including on scanned copies read with OCR.
Technical specifications
Parameters, limits, interfaces and conformance requirements, read into a requirements table with one row per requirement. Each row links to the clause that states it, so a conformance check cites its source.
Regulatory change and obligations registers
An obligations register lists what each rule requires, of whom and by when, one row per obligation. We read it from the regulation text, and every row links to the passage it came from.
| rule | Regulation SHO, Rule 201(b) |
|---|---|
| party | trading center |
| obligation | written policies and procedures reasonably designed to prevent a short sale at or below the national best bid |
| trigger | price ≤ 0.90 × prior close |
| duration | rest of trigger day + the following day |
| source | 17 CFR 242.201(b)(1), linked to its phrase |
| review | reviewer, time and source version |
When a rule changes
We compare the new version of the text with the one each row was read from. Rows tied to a changed passage go back to review, and each row keeps its source version and every change since.
Discuss an obligations registerRules as code: from text to model
Rules as code writes a rule's text as a model a program can check. One example: the short-sale price test in SEC Regulation SHO Rule 201(b). The same steps apply to any clause, obligation or requirement.
Source text, 17 CFR 242.201(b)(1)
A trading center shall establish, maintain, and enforce written policies and procedures reasonably designed to:
(i) Prevent the execution or display of a short sale order of a covered security at a price that is less than or equal to the current national best bid if the price of that covered security decreases by 10% or more from the covered security's closing price as determined by the listing market for the covered security as of the end of regular trading hours on the prior day; and
(ii) Impose the requirements of paragraph (b)(1)(i) of this section for the remainder of the day and the following day when a national best bid for the covered security is calculated and disseminated ...
Extracted record
| rule | Regulation SHO, Rule 201(b) |
|---|---|
| trigger | price ≤ 0.90 × prior close |
| baseline | listing market close, prior day |
| restriction | no short sale at or below national best bid |
| duration | rest of trigger day + the following day |
| exceptions | (b)(1)(iii)(A) and (B), extracted as separate records |
| source | 17 CFR 242.201(b)(1) |
Each field keeps a link to the exact phrase it came from. Point at a field or a phrase to see the pair.Each tag in the text names the field it became.
A stated property, then a check
Once the record is reviewed, we write down what the rule requires and test the model against it with property-based testing: generated cases search for one that breaks the stated property.
Property, stated in plain language and encoded in the model
Once the trigger fires, no short sale order executes or displays at or below the national best bid until the end of the following day, except as (b)(1)(iii) allows.
Boundary cases
Checks start where the text draws lines. A fall of exactly 10% triggers the restriction, because the rule says "10% or more". A trigger in the last minutes of the session carries the restriction through the next day.
Generated sequences
We generate sequences of quotes, orders and cancellations around the trigger and run each one through the model. Any sequence that breaks the property comes back as a counterexample, with the clause involved.
Recorded history
The same model can also run over the U.S. equity trade and quote history from 2018 that Rillor holds, to compare trigger and duration logic with recorded prices as well as generated cases.
The result is one of two outcomes. The property holds on the model, or the check returns a concrete sequence of events that breaks it, traced back to the clause in the source text.
Market rules as checkable models
We model how a trading system accepts, prioritizes, matches and cancels orders, and how scarce capacity is allocated. Then we state the properties it should hold and search for counterexamples.
What we test
- Priority rules such as price-time priority, where the best price fills first and orders at the same price fill in the order they arrived.
- Properties such as "every order shown in the book can execute" and "moving the price requires capital at risk".
- Counterexample search: sequences of orders and events that break a stated property.
- Timing and sequencing: order arrival, cancel races and stale state.
- Paths to manipulation or disruption through a venue's public interfaces.
- Repairs: proposed rule changes, checked so normal operation keeps working.
Where we apply it
- EquitiesOrder types, price-time priority, opening and closing auctions, Limit Up-Limit Down price bands and short-sale rules such as Rule 201.
- FuturesMatching and allocation rules, price limits and trading halts.
- Decentralized exchangesOn-chain order books and perpetual futures, including matching, liquidation and funding rules.
- Compute capacitySpot, preemptible and reserved pricing, interruption rules, reservation windows, and who gets capacity when it runs short. Rillor Compute Index
The same evidence rule in production
Two of our own systems hold every extracted fact to its source passage.
Grellum
Agentic legal research. Every citation in an answer must match the retrieved source passage, across the full corpus.
Every citation checked against the source passage it came from
grellum.comPublic-records platform
Archives public procurement records and regulation text, extracts document text, including OCR, and tracks changes, so every record links back to the document and version it came from.
Every record links to the document version it came from
Public-records intelligence
Extraction accuracy and human review
The LLM proposes each field. A reviewer accepts it before the record is used.
- What the reviewer seesEach proposed field beside the passage it came from, with the phrase highlighted.
- What the reviewer doesAccepts, corrects or rejects each field.
- What the record keepsThe reviewer, the time, the source version and every change since.
- When the source changesRecords that depend on the changed passage go back to review.
- Where corrections goInto the test set for the next extraction run, so a repeated error is caught before review.
- Documents from outside sourcesDocuments reach an extraction agent marked as data from outside, never as its instructions. Text planted in a document to steer the agent is indirect prompt injection, one of the attacks we test agents against. Agentic security
Common questions
How is this different from intelligent document processing?
Our work starts from text that sets rules, obligations or requirements. Every field keeps a link to the exact phrase it came from, and a person reviews every record before it is used. A rule can go one step further, into a model checked against stated properties.
Can you read scanned pages?
Yes. Scanned pages are read with OCR, and each field still links to its page and passage, as it does in a text document.
What happens when a regulation changes?
We compare the new version of the text with the one each record was read from. Records tied to a changed passage go back to review, and each record keeps its source version and every change since.
Research
The newest articles from our Insights library, each with a full research report and its sources.
- EU AI Act obligations as a register: providers and deployersProvider and deployer duties in Regulation (EU) 2024/1689, read into register rows by role, system scope and date, with each change from the Digital Omnibus traced.Document and regulation intelligence7 min readPDF report
- What court records show about fabricated citations in LLM outputCourt opinions and sanctions orders from 2023 to 2026 show what judges found in hallucinated legal citations and what they ordered in response.Document and regulation intelligence7 min readPDF report
- Extracting data from SEC filings: XBRL and textWhich parts of a 10-K, 10-Q or 8-K must carry XBRL tags, what the data.sec.gov APIs return, and which sections stay as text to extract.Document and regulation intelligence7 min readPDF report
Notices
SI outputs. Superintelligence (SI) outputs can be wrong. Important decisions should include human review.
Bring us the documents you rely on.
Send a sample of the rulebooks, regulations, contracts or filings you work from and the fields you need from them.