Privacy policy

Last updated 28 October 2026. Inferam is operated by InferamAI.

Inferam serves open AI models to developers and their coding agents. You send us source code and prompts; we treat them as yours and never keep them.

What we never store

• Prompts, messages, or system instructions you send to the API
• Completions, reasoning output, or tool calls the model produces
• Files, images, audio, or other media you upload

This content passes through our systems only to produce a response. It is not written to databases, logs, traces, analytics, or error reports. Prompt-cache state used to speed up repeated requests is held in memory on our inference servers, is never shared between accounts, and is discarded on restart.

We never train on your data

We do not use anything you send us, or anything the model returns to you, to train or tune models.

What we do store

For each API request, we record only this metadata:

• Account ID and API key ID
• Model name and timestamp
• Input, output, and cached token counts
• Time to first token and generation duration
• Response status code

We use it to show your usage dashboard, operate and secure the service, and plan capacity. Request metadata is deleted after 6 months.

We also store your email address (for your account, waitlist entry, and sign-in links), the API keys you create (as irreversible hashes only), and sign-in sessions.

Service providers

Inferam runs on Cloudflare (hosting, database, email delivery) and uses Grafana Cloud for operational monitoring. Monitoring receives metrics and request metadata only, never request content. Model inference runs on hardware operated by InferamAI.

Your choices

You can revoke API keys at any time. To delete your account and its metadata, or for any privacy question, email yash@inferam.ai.

Changes

If we change this policy in a way that reduces your privacy, we'll email you before it takes effect.