How AI Processing Works
Pending legal review. This page supplements the Privacy Policy and the PMA Agreement.
Where your material is kept. Your account, chats, uploaded documents and saved memory are stored on a server operated by TheLiving.AI. The website's pages are served by Vercel. When you use the site, your browser sends your requests, including chats and uploads, to our service through Cloudflare's network, which carries them to our server and carries our responses back.
What is sent to an AI model.
- To answer you, we send an AI model the text it needs: your message, plus related context we retrieve for it, such as earlier conversation or passages from your saved material.
- Chat models are reached through Venice (venice.ai).
- The search index of your saved material is built on our own server. That step sends nothing to an outside AI provider.
Your written words are not redacted.
- We do not remove names, email addresses, account details, case numbers or other information you type. They remain part of the text sent to the model.
- Private models: Venice states that the providers of Private models operate under contractually enforced zero data retention.
- Anonymized models: Venice states that it obscures identifying metadata about you. The upstream provider still receives the full contents of the prompt, including any identifying details written in it.
Account and billing information. Our own sign-in and billing records are handled as part of operating the service. Connection details, such as your IP address, reach the services that carry your request. Anything you write in a message, including account or billing details, is sent to the model as written.
Other services involved in running TheLiving.AI. Based on how the service is currently configured, these services handle member information:
- Venice processes the text sent to AI models, as described above.
- Vercel serves the website's pages.
- Cloudflare carries your requests and our responses between your browser and our server.
- Resend sends account emails, such as password resets. It receives your email address and the message we send.
- Stripe will process payments when paid checkout opens. It will receive your billing details. Paid checkout is not offered today, and we will update this page before it opens.
This page describes the services we have identified in our current setup. We will update it if that changes.
Venice's statements are Venice's. The descriptions of Private and Anonymized models come from Venice's privacy documentation (docs.venice.ai/overview/privacy). They are not a separate promise from TheLiving.AI, and they are not an agreement between you and Venice or its providers. We do not claim that trusted-execution or end-to-end-encrypted model options are used for your requests.
Memory ranking ("Jev"): enabled for members on October 1, 2026.
- Jev is enabled with daily call limits and a short response deadline. If a limit is reached or Jev is unavailable, Aethon uses the standard scoped retrieval ranking.
- What it does: when on, Jev helps choose which of your saved passages are most relevant to your message. It sends your message and a small number of candidate passages to Venice's
jev-latestmodel, which Venice labels Anonymized. The upstream provider receives that text in full. - What it can see: Jev only re-orders passages you are already permitted to see, from your own account and the case in use. It cannot give access to anything else.
Our direction. We intend to run every model on servers TheLiving.AI owns, so that member text never leaves our infrastructure. That is a roadmap goal, not how the platform works today.
Source: docs.venice.ai/overview/privacy
Appendix
AI models that can process member text
Labels are those shown in Venice's model catalog on 1 October 2026. They can change.
| Use | Model | Runs on | Venice label | Status |
|---|---|---|---|---|
| Chat for Free and Sovereign members | deepseek-v4-flash | Venice | Private | Active |
| Chat for Dragon and Sovereign Court members | deepseek-v4-flash-0731 | Venice | Private | Active |
| Memory ranking ("Jev") | jev-latest | Venice | Anonymized | Enabled with limits |
Image generation: availability is not confirmed, and we make no privacy-label claim for it. Voice: not offered to members.
Services
| Use | Service | What it receives | Status |
|---|---|---|---|
| Storage and the search index of your saved material | Our server | Your account, chats, documents and memory | Active |
| Website pages | Vercel | Requests for the site's pages | Active |
| Delivery between your browser and our server | Cloudflare | Your requests and our responses | Active |
| Account email (password reset) | Resend | Your email address and the email we send | Active for password reset |
| Payments | Stripe | Billing details | Not offered yet; this page will be updated before checkout opens |
Roadmap
- Sovereign infrastructure: run every model on TheLiving.AI-owned servers, so that no member text leaves our infrastructure. This is a goal, not current practice.
- Member voice: not offered today. This page will be updated before voice is offered.