Data Architecture & Privacy

Your case data stays
exactly where it belongs.

OptoJury is built on API-only inference — no data is stored, logged, or used to train any model. Every query is stateless. Every document is yours alone.

How the architecture works
01
You submit a query

Your case narrative, document, or prompt is sent over encrypted TLS to the inference API. It never touches a shared database or logging layer.

02
Stateless inference runs

The model processes your query in memory only. No session is persisted. No input is retained after the response is returned. There is no "your data" on our servers.

03
Output returns to you

The diagnostic report is delivered directly to your session. Nothing is cached, cross-referenced with other firms, or retained for model improvement.

Privacy guarantees
Never used for training

OptoJury uses commercial API providers under data processing agreements that explicitly prohibit the use of API inputs for model training. Your privileged work product does not improve anyone else's model.

No query logging

Queries are stateless by architecture, not by policy. There is no logging infrastructure to disable — each request is processed and discarded. We cannot retrieve what you submitted last week because it does not exist.

Firm isolation

No shared context windows, no cross-firm retrieval, no aggregated data. Your queries are never visible to other users. The system has no memory of your prior sessions.

Attorney-client privilege preserved

Because no third party retains your data, the disclosure argument that undermines privilege in cloud AI tools does not apply here. API-only, stateless processing is the architecturally correct answer to Rule 1.6.

OptoJury vs. general AI tools
Risk Factor ChatGPT / Claude.ai (direct) OptoJury
Query retention Stored for 30 days (default); may be reviewed by provider staff
Stateless — zero retention
Used for training Opt-out required; API-only exempts but consumer apps do not
API DPA prohibits training use
Privilege risk Third-party disclosure argument applies under most bar guidance
No third-party retention — disclosure argument does not arise
Cross-firm contamination Shared infrastructure; context leakage documented in consumer tiers
Isolated per-request, no shared state
Data residency Jurisdiction varies; cross-border transfer may occur without notice; GDPR/CCPA exposure unresolved
U.S.-based API endpoints; no cross-border data transfer by default; residency commitments available on request

Built for
the profession's rules,
not around them.

Stateless, isolated, API-only. The architecture is the answer — no policies to toggle, no opt-outs to manage.

Access OptoJury
Governing Standard

ABA Model Rule 1.6(c)
"A lawyer shall make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client."

Stateless, API-only processing eliminates the retention vector that makes consumer AI tools non-compliant under most state bar opinions on cloud services and generative AI.