Legal Help Commons

Standards and Specifications

Open specifications for what legal help tools should do, how well they should do it, and how to organize the data they depend on. Developed by practitioner working groups, open for anyone to build to.

Functional Agendas and Conformance Standards Standards by Tool Reference Standards and Taxonomies
Standards by Workflow

What Legal Help Teams Want Built, and How Good It Needs to Be

Each workflow produces a functional agenda (what the system must do) and conformance standards (how well it must do it). Some also have test suites and other evaluation resources. All documents are drafts open for review and comment.

Voice AI Intake and Triage

Help people who call a legal help hotline get screened, triaged, and routed to the right service through a voice-based AI assistant.

Draft for review Functional Agenda What the system must do at every stage of a call, from greeting through handoff. Draft for review Conformance Standards Five gates the system must pass: performance, usability, safety, onboarding, sustainability. Draft for review Test Suite Realistic caller scenarios with scoring rubrics and expected outcomes.
See the full Voice AI Intake package →

Eviction Notice Defect Spotting

Help a legal aid worker read an eviction notice, flag statutory defects, and turn those flags into defenses and a court response.

Draft for review Functional Agenda What the system must do from reading the notice through generating a response.
In development Conformance Standards Detection accuracy, false positive rate, evidence quality, and safety gates.
See the full Eviction Defense package →

Brief Help Legal Q&A

Help people who describe a legal problem get accurate, jurisdiction-specific guidance, next steps, and connections to services.

Draft for review Functional Agenda What the system must do from understanding the question through delivering an answer with citations and next steps.
In development Conformance Standards Accuracy, jurisdiction sensitivity, source fidelity, honest uncertainty, and safety.
Standards by Tool

The Shared Pieces Many Tools Need

Common infrastructure tools are the components that several workflows depend on. A tool standard says what the component must do, and how to score a system against it, so a team can choose among existing options instead of building another one. Browse all of them on the common infrastructure tools page.

PII Masker

Find and remove personal information from legal documents before they are stored, logged, shared, or sent to a model.

Draft for review Conformance Standard Three binary gates, five scored criteria, and the entity tiers a system is scored against.
In development Benchmark Set Three hundred samples, half clean text and half scanned, with human-labeled spans and severity tiers.
See the full PII Masker package →

Legal Issue Classifier

Take text describing a person's situation and return the codes from a shared legal issue taxonomy that say what kind of problem it is.

Draft for review Functional Agenda What a classifier must do, from accepting a three-word phrase through returning a code path with confidence and urgency. Draft for review Conformance Standard Binary gates a system cannot fail, scored dimensions under the five Commons gates, and a mapping from score to what the system may do.
In development Test Sets Eight datasets the suite runs against. Most do not exist yet, and building them is what the cohort is forming to do.
See the full Legal Issue Classifier package →

Tool Standards in Development

Standards for shared infrastructure tools that multiple workflows depend on.

Coming

Document Reader

Read court filings, notices, and legal documents via OCR and extraction.

Coming

Eligibility Checker

Determine service qualification against income, geographic, and case-type rules.

Want to help define an agenda? Each one is developed by a practitioner working group. Join a working group to help shape what gets built and how it gets tested.
Reference Standards and Taxonomies

The Shared Infrastructure Underneath

These standards define how legal help data is organized, how legal issues are classified, and what tasks providers and users perform. Every workflow and tool in the Commons depends on them.

Legal Help Knowledge Standard v0.1 Draft

How to organize information about legal help organizations, the services they offer, and the content available to the public. Shared field names, categories, and values so one organization's data can be found and used by another. Use this to structure your articles, guides, FAQs, and service directories so AI tools can use them reliably.

Read the Knowledge Standard →
Taxonomy Crosswalks → Content Index Standard →

LIST: Legal Issues Taxonomy

A standardized vocabulary of the specific legal problems people experience: eviction, wage theft, child custody, debt collection, and hundreds more. Used to classify what kind of help someone needs and to connect them with the right resources and services.

Browse the LIST taxonomy at taxonomy.legal ↗

Legal Help Task Taxonomy

A catalog of the specific activities that providers and users perform when dealing with a legal problem: answering questions, classifying issues, reading documents, calculating deadlines, preparing filings, and more. Use this to identify where AI can help and what to measure.

Browse the Task Taxonomy on JusticeBench ↗
How these standards work together. The Knowledge Standard defines how legal help content is organized. The LIST taxonomy classifies the legal problems people face. The Task Taxonomy catalogs what providers and users do. The functional agendas define what tools should do to help with those tasks. The conformance standards test whether the tools do it well enough. Each layer builds on the others.
Open specifications, not vendor endorsements. The Commons does not approve or rank products. It defines what the field needs, publishes the specification, and tests whether implementations meet it. Any builder, vendor, or organization can build to the specification. Conformance results are published on JusticeBench.