Zero Retention
Data Handling at abliteration.ai
Prompts and completions are processed transiently and never stored by default. We retain only what we need for billing and reliability.
Last updated: January 22, 2026
What We Store
Only operational data needed to run the service.
- Token counts: How many tokens each request consumed.
- Timestamps: When requests were made for usage dashboards.
- Error codes: HTTP status codes for reliability monitoring.
- Billing metadata: Invoice references, payment confirmations.
What We Do Not Store
Your content never hits disk.
- Prompts: Your input text is processed transiently in memory.
- Completions: Model outputs are streamed directly to you.
- Images: Vision inputs are processed and discarded.
- Conversation history: We do not maintain server-side chat logs.
Never Used for Training
Your data stays yours.
- Prompts and completions are never used to train or fine-tune models.
- We do not share your content with third parties for any purpose.
- Aggregated, de-identified metrics may be used for reliability improvements only.
FAQ
- Do you log my prompts for debugging?
- No. Prompts are processed in memory and never written to disk or logs. Error codes and timestamps are logged, but not content.
- What about the Policy Gateway?
- Policy Gateway stores rule configurations, enforcement decisions, and audit metadata. It does not store prompt or completion content. See Policy Gateway Security.
- Can I request deletion of my data?
- Yes. Contact help@abliteration.ai to request deletion of your account and associated billing records. Legally required records may be retained.
- How long is telemetry retained?
- Operational telemetry (token counts, timestamps, error codes) is retained for billing reconciliation and fraud prevention, typically for the duration of your account plus any legally required period.