See the flow

For trial counsel & hearing teams · Virginia & District of Columbia first

The right authority when
the objection lands

MatterBench is a second screen at counsel table for trial and evidentiary hearings. Capture the exchange that just occurred and surface ranked authorities, pinpoint citations, and matter-file references while the record is still open— analysis grounded in disclosed sources, for counsel’s independent judgment.

  • Live recommendation cards during argument, cross, and objections
  • Litigation pack first; optional public reporters when policy and network allow
  • Local-first processing—privilege posture stays with the firm laptop
How it works Built for trust

Research aid only · Not legal advice · Counsel verifies before reliance in open court

How a hearing day works

Three steps your trial team already runs—wired for live authority support without a separate research sprint mid-hearing.

  1. Open the matter Ingest the litigation pack—pleadings, orders, authority memoranda, exhibit index, and working notes—then open the hearing session for the day.
  2. Capture the record or pin the issue Note the exchange that just occurred, or pin a discrete issue (e.g., confrontation, Va. R. Evid. 2:403, present-sense impression). Research begins immediately against the pack and, if allowed, public reporters.
  3. Work the live authority queue Cards rank analysis, key citation, and provenance. Follow up, dismiss noise, and export a post-hearing trail for the file and the team.

During argument

Authorities land as issues arise—not only at recess. Pin anything detection misses so the queue stays under counsel’s control.

At recess

Same queue, more time. Review pinpoint quotes, run follow-up research, and export an annotated trail for the record and trial team.

Counsel’s independent judgment

Every card is research support. Confirm authorities against primary sources before reliance in open court or in a written submission.

Built for the people at counsel table

Trial counsel

Second-screen support while you examine and argue. Highest-ranked pack authorities first; expanded public-law research when your matter policy permits.

Second chair & associates

Capture the exchange, pin discrete issues, and hand lead counsel a ranked authority queue instead of a fragmented chat thread.

Paralegals & trial support

Prepare the matter pack once. Operate the hearing-day view. Export the trail for the case file after the sitting concludes.

Not another desk research tool

Office legal AI
Deposition tools
MatterBench
When you use it
Briefing & drafting
Deposition room
Trial & evidentiary hearings
What you get
Memos & drafts
Witness inconsistencies
Live authority cards & pinpoints
Research posture
Cloud corpus / chat
Case-file compare
Matter pack first, local-first, typed provenance

Built so you can rely on the queue’s discipline

Hallucinated reporter cites and silent overconfidence are malpractice risk, not a product feature. MatterBench treats those failure modes as constraints on what may leave the queue as “ready.”

Research aid. Counsel decides.

MatterBench does not practice law and is not a substitute for independent legal judgment. Confirm every authority before reliance in open court or in a filing. Capture is designed for counsel-table notes—not a product for surreptitious full-room recording.

Standalone product. Not affiliated with “Second Chair,” CoCounsel, NexLaw, or similar brands.

Technical detail

For product, eng, and ops reviewers. Counsel teams can stop above.

System path

Capture → session issues → research fabric → analysis → live queue and trail.

Capture STT · segments Session core issues · jobs Research fabric pack · public · comm. Analysis agents tier-1 · tier-2 Live queue cards · cites · trail Local-first · matter isolation · live card delivery during argument
End-to-end path from counsel table to live authority cards

Two-tier research pipeline

Tier-1 is optimized for recess latency and offline courtrooms. Tier-2 expands the universe only when compute policy and connectivity allow—and labels that choice in the UI.

Issue hypothesis TIER 1 · FAST Matter pack FTS + RRF authority seeds · pleadings · local rules target: seconds · offline OK TIER 2 · DEEP Public law + optional commercial CourtListener · jdx court filters · cache async · policy + network gated Card synthesis claims · no-invention eyecite · CL lookup Queue ready · partial
Tier-1 stays pack-local for speed and offline use; tier-2 supersedes when richer sources land

Issue detection

Heuristic overlap of rolling transcript tokens with claims_or_charges and authority-seed titles. Confidence floor 0.4; 60s debounce; counsel pin always wins ranking budget.

Hybrid retrieval

SQLite FTS5 per-matter chunks fused with Reciprocal Rank Fusion (k=60). Note-class artifacts never surface as legal sources—ranking bias only.

Public law client

CourtListener REST v4: Authorization: Token …, GET /search/?type=o, soft-fail on 401/403/429, matter-scoped TTL cache, citation-lookup guardrail.

Compute policy

Per-matter flags: local default, allow public law, allow commercial, allow remote models. Session data-flow chips reflect live state in the recess UI.

Architecture that keeps privilege local

Single-laptop process for v1: FastAPI session core, SQLite matter store, pluggable backends, Vite review shell. No runtime dependency on Cerid AI or multi-tenant cloud for the courtroom path.

MATTER A chunks · FTS cards · events hearing sessions · segments · issues MATTER B chunks · FTS cards · events no shared vector space in v1 ✕ Hard isolation — every query is matter_id-scoped
Cross-matter retrieval is structurally impossible in the default store

Primary API surface

POST /matterscreate matter + compute policy
POST /matters/{id}/ingestlitigation pack artifacts
POST /matters/{id}/hearingsstart hearing session
POST /hearings/{id}/segmentsSTT → detect → jobs
GET /hearings/{id}/queueactive recommendation cards
POST /cards/{id}/followupfollow-up research job
GET /hearings/{id}/trailannotated audit bundle
GET /hearings/{id}/exportMarkdown or JSON download

No-invention gate

Legal AI fails on invented citations and silent overconfidence. Failure modes are product features: strip, flag, degrade—never invent.

Draft card LLM or determ. eyecite parse cite spans Known-set check hits · sources only Strip · flag · rebind adverse: stripped_unverified_citation optional CL citation-lookup
No-invention gate: fabricated reporter cites never leave the pipeline as “ready” analysis
  • Sources + analysis only — card schema and synthesizer prompts emit authorities and analysis for counsel review, not bench advocacy copy.
  • Provenance types — every source is matter_pack · public_law · commercial · transcript.
  • Coverage honesty — cards report pack_only · partial · full; researching never looks final.
  • Local-first compute — optional remote models/search labeled per matter; offline tier-1 remains usable.
  • Eval gates — golden in-pack recall + zero fabricated cites in the automated suite.

Research aid. Counsel decides.

MatterBench does not practice law. Confirm authorities against primary sources before reliance. Counsel remains responsible for recording compliance—defaults assume counsel-table capture, not full-room recording.

Standalone product. Not affiliated with “Second Chair,” CoCounsel, NexLaw, or similar brands.

Jurisdiction pack: va_dc

First-class data package—not hard-coded ifs. Binding hierarchy and public backend weights travel with the matter.

STATE · VA Supreme Court of Virginia binding_state CL id: va STATE · DC D.C. Court of Appeals binding_state CL id: dc FEDERAL FOOTPRINT vaed · vawd · dcd ca4 · cadc · scotus preferred_court_ids → CL type=o search
Jurisdiction pack supplies court ids and source weights for ranking

Source weights (v0)

  • matter_pack
    2.0
  • binding_local
    1.5
  • binding_federal
    1.3
  • persuasive
    0.8
  • note (bias only)
    0.0

Public sources may miss materials commercial databases cover. Cards disclose partial coverage when backends fail or are offline.

Technical stack

Concrete components shipping on the Moment A path—not a vaporware diagram.

Runtime

  • Python 3.12+ · FastAPI session core
  • SQLite + FTS5 matter isolation
  • Pydantic v2 domain models
  • RRF fusion (Apache-licensed extract)
  • httpx · eyecite · pypdf

Capture & agents

  • faster-whisper batch STT
  • Streaming STT adapter interface
  • Capture CLI (demo / file / mic)
  • Deterministic synthesizer default
  • Optional OpenAI-compatible LLM cards

Surfaces & quality

  • Vite + React live queue UI
  • Trail panel + MD/JSON export
  • Golden evals: recall + no-invention
  • CI: ruff · pytest · coupling ban
  • Product site: matterbench.io
compute_policy · example 
{
  "local_default": true,
  "allow_remote_models": false,
  "allow_public_law": true,
  "allow_commercial": false,
  "commercial_backends": []
}

Research aid. Counsel decides.

Confirm authorities against primary sources before reliance in open court. Counsel remains responsible for recording compliance at the table.

How a hearing day works Built for trust

Local-first · Matter isolation · Jurisdiction packs without forking the core