Skip to content

Category: Ember Legal · Legal Technology

Bilingual legal intelligence for the PRC–international corridor — every citation survives verification.

Ember Legal retrieves from a real, PRC-dominant legal corpus, reasons across legal systems side by side, and verifies every quoted authority verbatim against the retrieved source before it is shown — abstaining when unsupported. The platform is in design-partner access: registration is invite-only while it is hardened with its first users.

Built for

  • Law firms
  • In-house legal teams
  • Cross-border arbitration teams
Status
Live
Ledger step
Ledger step 4 of 4
Family
Ember family
Access
Invite only
Listed
4 / 4
Citation verification
Verbatim
Never fabricate
Abstain
Native bilingual workflow
EN + ZH
Design-partner access
Invite-only

See it working

A real capture of the live product, not a mock.

The Ember Legal home screen, where PRC and international legal sources are retrieved and compared side by side. Captured 18 September 2026
Ember Legal home page showing the bilingual legal research and jurisdiction comparison tool.

What this page can show today

  • The register marks this track Live.
  • Capabilities below are the published feature list for this track.

Limits of this page

  • Access is invite-only. A public URL here is not an open signup.
  • This page does not publish financials, headcount, or client names.

Key features

  1. 01

    Verified citations, by design

    Every quoted authority is checked verbatim against its source before it is shown. Unsupported output abstains instead of fabricating, the failure mode that disqualifies generic assistants in a filing.

  2. 02

    A real PRC-heavy corpus

    644K legal documents and 1.04M retrievable chunks, 96.5% PRC-jurisdiction, enriched with deterministically mined award and court-level analytics.

  3. 03

    A full case workbench

    A 23-tab litigation workbench across 63 live modules, covering the case lifecycle from intake and evidence to drafting and review.

  4. 04

    Two legal systems, two languages

    Reasoning spans PRC and US legal systems with fully bilingual EN/中文 output, built for the cross-border matter.

Technology stack

Next.js · FastAPI · PostgreSQL · Ember Reason · Ember Guard

Related platforms

Family on the register: Ember family Same family first, then tracks at the same lifecycle stage.

  • Same family · Live

    Ember AI

    Foundation Models

    Public product site

  • Same family · Beta

    Ember Live

    Real-Time Model Surface

    Invite only

  • Same family · Beta

    Ember Desktop

    AI Work Operating System

    This site only

  • Same family · Beta

    Ember Agent OS

    Multi-Agent Operations

    This site only

This cell on the investor register

Status and family write into the /investors URL. The holdings table lists every public track in that cell.

Open Live Ember family holdings
4 / 4

JSON fields are copied from this row in src/data/platforms.ts. No financials.

How this track works

Audience on the left, the published Ember stack in the middle, capabilities on the right. Every node is copied from this track in the public register.

How this track moves work from its published audience through Coreidea and Ember into listed capabilities.

Inputs

Coreidea and Ember

Outputs

Select a node to inspect the published role and capability.

Cross-border legal work is growing faster than its tools.

PRC–international disputes and contracts are expanding, but the tooling remains monolingual, single-system — and generic assistants fabricate citations.

  • Monolingual, single-system tools

    Mainstream legal tech serves one language and one legal system at a time. The cross-border matter lives in both.

  • Fabricated citations are disqualifying

    A generic assistant that invents a case citation is not an inconvenience in legal work — in a filing it is disqualifying.

  • Opaque reasoning cannot be relied on

    Legal conclusions need a visible chain from source to statement. Black-box answers cannot carry professional responsibility.

Honest by design — retrieve, generate, verify.

The platform is built around one loop: retrieve from a real corpus, reason bilingually, verify verbatim, abstain when unsupported.

  • Verbatim citation verification

    Every quote is checked against its source document before display. What you see has been matched to the text it cites.

  • Abstention over fabrication

    When the corpus does not support a claim, the platform says so and abstains — it does not fill the gap with fluent invention.

  • A glass-box corpus

    The platform publishes its own coverage per jurisdiction with a date — China-dominant today, with active English-language backfill from license-screened sources.

  • Bilingual across two legal systems

    PRC-law retrieval combined with common-law reasoning, fully bilingual EN/中文 — built for the corridor where both apply.

From question to verified answer.

  1. Retrieve

    The question is answered from a real, per-jurisdiction corpus whose live coverage the platform publishes itself — not from a model memory.

  2. Reason

    Analysis runs bilingually across PRC and US legal frames, matched to the posture of the matter.

  3. Verify

    Every citation is verified verbatim against its source before it reaches the screen.

  4. Deliver or abstain

    Supported conclusions are delivered with their sources; unsupported ones are declined, visibly.

What runs underneath.

  • The verified pipeline

    A working platform, not a demo: intake, retrieval, bilingual drafting, verification and tracked filing run as one pipeline — with a gate before anything leaves.

  • A defined matter boundary

    Matter documents, drafts and review notes stay in per-account case files — never used to train shared models, with enterprise isolation options verified per engagement.

  • Coverage published, not sloganeered

    A live library dashboard shows per-jurisdiction document counts and decision-date spans as a dated snapshot — coverage is a moving number, not a marketing total.

  • Ember model family inside

    Reasoning and verification run on Coreidea’s own Ember engines under the group’s privacy and residency practices.

Built for the corridor matter.

  • Cross-border disputes

    PRC–international litigation and arbitration where authorities from different systems must be found, read and cited correctly.

  • Contracts across systems

    Drafting and reviewing agreements that span PRC law and other jurisdictions, with cross-version enforcement differences flagged.

  • Cross-system research

    Per-jurisdiction retrieval with a side-by-side dimensions table — how different jurisdictions treat the same doctrine, without cross-contamination.

Questions, answered.

Is Ember Legal publicly available?

The platform is in design-partner access: registration is invite-only while it is hardened with its first users. The public investor overview linked on this page describes it in full.

Does it fabricate citations like generic assistants?

Its defining design choice is the opposite: every citation is verified verbatim against the source before display, and unsupported output abstains.

What is in the corpus?

A PRC-dominant legal corpus with active, license-screened English-language backfill. The platform publishes per-jurisdiction counts as a dated dashboard snapshot rather than quoting a marketing total.

Where does client data live?

Matter data stays inside the platform under the group data-segregation and privacy practices; library holdings are private by architecture.

Read the investor overview.

The public investor overview covers market, product, model and roadmap. For engagements or the due-diligence pack, contact the group directly.