Third-party services this product relies on
- JurisGPT manages account sign-in and revocable sessions through its own backend.
- Supabase (hosted PostgreSQL) is the application database. The current schema defines tables for user profiles, companies, founders, legal matters, documents, lawyer reviews and legal preferences.
- Questions submitted to the assistant are sent to a third-party large language model provider — OpenAI or Anthropic, depending on how the deployment is configured — which generates the answer text.
- Retrieval runs against a local vector index built from public Indian legal documents.
Telemetry and analytics
- No third-party authentication SDK is bundled in the frontend.
- The vector store client is configured with anonymised telemetry switched off.
- No product analytics or advertising SDK is bundled in the frontend: the project has no Google Analytics, PostHog, Mixpanel, Segment or Vercel Analytics dependency.
Request logging
The API records each request in a bounded, in-memory audit log: request id, timestamp, client IP, HTTP method, path, status code, response time and, for signed-in requests, user id and email. Values that look like secrets are redacted. This log lives in process memory and is not written to the database.
The public demo
The demo requires no account. Questions are proxied through this site's server to the backend; the demo does not ask for or attach an identity.
NOT DECIDED OR DOCUMENTED YET
Deliberately left blank rather than filled with plausible-sounding text. Each item needs a decision and legal review before it can appear here.
- What is retained, where, and for how long
- How to request access, correction, export or deletion
- The lawful basis for each kind of processing, and how consent is captured
- A named contact route for privacy questions or grievances
- A complete sub-processor list and where each one stores data
- Whether and how data crosses borders