Federal court data is some of the most valuable structured information in the United States — and some of the hardest to access programmatically. This guide explains how the federal court system is organized, what PACER is, how CM/ECF works, and why this data matters for AI agents operating in legal, financial, and compliance workflows.

How the federal court system is organized

The United States federal court system has three tiers.

District Courts are the trial courts of the federal system. There are 94 federal judicial districts across all 50 states, the District of Columbia, and US territories. District courts handle civil cases, criminal cases, and bankruptcy cases. When a company is sued in federal court, when a federal crime is prosecuted, or when a debtor files for bankruptcy protection, those cases originate in a district court.

Courts of Appeals (also called circuit courts) hear appeals from district courts. There are 13 circuit courts organized by geography — the First Circuit covers New England, the Ninth Circuit covers the western states, and so on. The Supreme Court sits above all of them.

Specialized Courts include the US Bankruptcy Courts (which technically operate as units of the district courts), the Court of International Trade, the Court of Federal Claims, and others.

For most legal, financial, and compliance AI workflows, district courts and bankruptcy courts are the primary targets. These are where the activity is — where lawsuits are filed, judgments entered, and bankruptcy petitions submitted.

The highest-volume federal courts

Not all federal courts are equal in terms of case volume. A handful of jurisdictions handle a disproportionate share of significant litigation and bankruptcy cases.

Southern District of New York

The most prominent federal court in the US. Securities litigation, financial fraud, complex commercial disputes involving major institutions.

court_code: nysd

District of Delaware

Preferred jurisdiction for major corporate bankruptcy filings. The busiest Chapter 11 court in the country.

court_code: deb

Northern District of California

Primary jurisdiction for technology and intellectual property litigation. Covers Silicon Valley.

court_code: cand

Central District of California

Largest federal district court by case volume. Covers Los Angeles and surrounding area.

court_code: cacd

Northern District of Illinois

Major commercial litigation hub covering Chicago. Significant financial and corporate disputes.

court_code: ilnd

Southern District of Texas

Growing jurisdiction for energy litigation and corporate bankruptcy filings.

court_code: txsd

Eastern District of New York

Covers Brooklyn, Queens, Long Island, and Staten Island. Significant criminal and civil cases.

court_code: nyed

District of New Jersey

Major pharmaceutical and healthcare litigation hub.

court_code: njd

What is PACER?

PACER — Public Access to Court Electronic Records — is the federal judiciary's official system for public access to court documents. Created in the 1990s and maintained by the Administrative Office of the US Courts, PACER is the authoritative source for federal court docket information.

Through PACER, anyone with a registered account can:

  • Search for cases by party name, case number, or attorney
  • View docket sheets showing all filings in a case
  • Purchase and download individual documents filed in cases

PACER charges $0.10 per page for most content, with a quarterly cap of $30 per account beyond which fees are waived. Commercial users — including businesses that access PACER programmatically — pay these fees at volume.

PACER has significant limitations that create the opportunity DocketLayer addresses:

No native real-time monitoring. PACER does not provide webhooks, push notifications, or change detection. To know if anything has changed in a case, you must query PACER repeatedly and compare results manually.

Fragmented access. Each federal court runs its own instance of CM/ECF. While PACER provides a unified login, the underlying data is court-specific and inconsistently structured.

No agent-native interface. PACER was designed for human attorneys and researchers. It has no API designed for programmatic consumption by AI agents.

What is CM/ECF?

CM/ECF stands for Case Management/Electronic Case Files. It is the software system the federal courts use to manage and store case information. When an attorney files a motion, when a judge issues an order, or when a party submits a document in a federal case, it goes into CM/ECF.

Every federal district and bankruptcy court runs CM/ECF. The system has been in place since the early 2000s and contains hundreds of millions of documents. From a developer's perspective, three things matter most:

Each court has its own instance. There is no single CM/ECF API that covers all federal courts. Each of the 94 district courts and 90 bankruptcy courts runs its own instance, with its own URL, its own login, and subtle differences in how data is structured and returned.

Access requires PACER credentials. Accessing CM/ECF programmatically requires a valid PACER account. Session management — authenticating, maintaining tokens, handling expiry — is a non-trivial engineering challenge that must be solved before any data can be retrieved.

Data is inconsistently formatted. While CM/ECF is nominally the same system across courts, different courts have customized it over the years. A docket entry in SDNY may be structured differently from the same type of entry in Delaware. Normalization — converting inconsistent raw data into a clean, uniform schema — is the core technical challenge in building any system that aggregates federal court data.

How DocketLayer solves these problems

DocketLayer sits between PACER/CM/ECF and the AI agents that need federal court data. It solves each limitation above:

Real-time monitoring with change detection. DocketLayer polls covered courts continuously and maintains a normalized cache of docket data. When an agent queries DocketLayer with a last_checked timestamp, DocketLayer compares the current cache state against that timestamp and returns only what is new.

Unified access across courts. Agents submit a case ID and a court code and receive a consistent, structured JSON response regardless of which court the case is in. The fragmentation of CM/ECF is invisible to the caller.

Agent-native interface. Payment is embedded in the HTTP request via the x402 protocol. No accounts, no subscriptions, no human steps. The payment is the authentication.

AI-powered normalization. DocketLayer uses an AI normalization layer to convert raw CM/ECF data into a clean, consistent schema across courts. This is what makes uniform coverage possible across courts with different data structures.

Types of federal court filings

Understanding what gets filed in federal court clarifies the use cases for automated monitoring.

Complaints

Initiate civil cases. When a company is sued, a complaint is the first document filed. Monitoring for complaints naming specific parties is a key use case for financial and compliance agents.

Motions

Formal requests asking the court to take a specific action — to dismiss a case, compel discovery, or grant summary judgment. Monitoring motion types helps agents track case progress.

Orders

The court's rulings on motions and other matters. A summary judgment order, preliminary injunction, or dismissal order can dramatically change the status of a case. Often the most important filings to monitor.

Judgments

The final outcome of cases. A judgment against a defendant creates an enforceable legal obligation — relevant to debt collection, insurance, and financial risk workflows.

Bankruptcy petitions

Initiate bankruptcy cases. Chapter 7 liquidations, Chapter 11 reorganizations, and Chapter 13 individual restructurings. Monitoring for bankruptcy filings by specific debtors is critical for creditors, insurers, and lenders.

Notices of appeal

Signal that a losing party is challenging a decision. Monitoring for appeals helps agents track whether adverse outcomes are final or still subject to reversal.

Federal case number format

Federal case numbers follow a standard format that encodes information about the case:

[division]:[year]-[type]-[number]

# Division — court division number, usually 1, 2, or 3
# Year — two-digit year the case was filed
# Type — cv (civil), cr (criminal), bk (bankruptcy), mc (miscellaneous)
# Number — sequential case number assigned by the court

1:24-cv-01234 # civil case, 2024, Southern District of New York
1:22-bk-11068 # bankruptcy, 2022, District of Delaware
5:20-cv-03010 # civil case, 2020, Northern District of California

DocketLayer accepts case IDs in this standard format along with a court code identifying which court to query. See the complete case number format reference for edge cases and variations across court types.

Why federal court data matters for AI agents

Federal court data is a leading indicator for a wide range of downstream events that matter to financial, legal, and compliance workflows.

Credit and lending risk. A bankruptcy filing or a large judgment against a borrower materially changes their creditworthiness. Agents monitoring loan portfolios can trigger risk reviews the moment relevant court activity appears.

Debt collection. Knowing when a judgment is entered, when a debtor files for bankruptcy protection, or when a case is dismissed is essential for collection workflow management. Agents can adjust strategies automatically based on case status changes.

Insurance. Claims-related litigation, coverage disputes, and judgments all flow through federal courts. Agents monitoring insured entities can detect litigation activity before it reaches a coverage decision.

Compliance and regulatory monitoring. Enforcement actions, consent decrees, and regulatory litigation are public record in federal court. Agents monitoring counterparties for compliance risk can detect problems earlier than traditional methods.

M&A and due diligence. Litigation exposure is a material factor in acquisition decisions. Agents performing due diligence can monitor federal court activity against target companies continuously.

Legal practice management. Law firms managing large case portfolios need to track activity across dozens or hundreds of active matters simultaneously. Agents can handle this monitoring load automatically.

Further reading