Rule 1.6 compliant by design

AI That Never Sees Your Clients' Data

Lethara strips and tokenizes client PII before any prompt reaches an external model. Move at the speed of AI without a breach, a bar complaint, or a second thought.

Subpoenas cannot threaten your attorney-client privileges when we seal your paralegal prompts.
Book a walkthroughSee how it works
New prompt
Vaulted
Draft a demand letter for [CLIENT_0427] regarding the collision on [DATE_0931] at [ADDRESS_0112]. Reference the policy held with [INSURER_0048] and request settlement within 30 days.
4 PII tokens secured · mapping stays local
How it works

Sensitive data is forgotten before it ever leaves the firm

The name comes from the River Lethe, the water of forgetting. Three steps stand between your client's data and any external model.

STEP 01

Intercept

Every prompt is caught at the membrane. Lethara scans for names, addresses, case numbers, and any identifier that could trace back to a client.

STEP 02

Tokenize

Each piece of PII is swapped for a reversible token. The AI sees structure and intent, never the underlying identity.

STEP 03

Restore

The response returns and Lethara re-inserts the real details locally, with a key only your firm holds. Nothing sensitive ever crossed the wire.

Features

Privacy you can hand to a client

Compliance-grade infrastructure that disappears into your workflow.

Local key vault

The token-to-identity mapping stays inside Lethara's protected workflow. The AI provider receives only tokens, so it cannot read client identities if subpoenaed.

Managed AI backend

Lethara handles the model connection behind the scenes, sending only tokenized prompts through the API layer we operate.

Audit-ready logs

Every interception is timestamped and exportable, creating a defensible paper trail.

Matter-aware safeguards

Legal prompts are checked for client identifiers before each model request, reducing exposure at the point of use.

Built for small firms

Up and running in an afternoon. No security team, no six-figure contract, no IT department required.

Real-time leak guard

Lethara replaces detected identifiers with tokens before AI send, and records a sanitized safety proof of what left the workspace.

The obligation

Model Rule 1.6 doesn't bend for new tools

A lawyer's duty of confidentiality applies the moment a prompt is typed. Lethara is the layer that lets you say yes to AI and still keep your oath.

Client data never reaches the AI providerIdentifiers are removed pre-flight before a model sees the prompt.
External models see tokensReversal happens after the model response returns, before your team reads it.
Defensible by recordExportable logs demonstrate compliance to a client or a board.
"A lawyer shall not reveal information relating to the representation of a client unless the client gives informed consent."
ABA Model Rule 1.6(a) - the line Lethara holds
Pricing

Simple seat pricing for small firms

ABA Model Rule 1.6 requires lawyers to make reasonable efforts to prevent inadvertent or unauthorized disclosure of client information. Lethara helps firms operationalize that duty with AI-facing confidentiality controls, audit evidence, and pricing built for law-firm buying patterns.

Monthly total
$395$79 / seat
Annual total
$4740$79 / seat × 12 months
Pilot offer · limited availability
$200/mo$40 / seat, 90-day term
  • ABA Rule 1.6 compliant workflow
  • Audit-ready session logging
Book a walkthrough
Book a walkthrough

See it handle a real client matter

Book a 20-minute demo. We'll show PII sanitization live on a sample client file and answer your compliance questions.