Privacy Policy

Last updated 31 July 2026

Scope

This policy covers the NeuralAccel inference API at api.neuralaccel.com, which serves open-weights language models to API clients and routing platforms. It also covers this website.

What we receive

When you send an inference request we receive the prompt content, model parameters, and standard connection metadata: IP address, timestamp, and request size.

What we retain

We do not retain prompt or completion content. Request and response bodies exist in memory only for the duration of the request. They are not written to disk and not written to any log.

We retain operational metadata only: timestamp, HTTP status, latency, token counts, and model name. This is what we need to bill accurately, measure reliability, and detect abuse. It is deleted after 30 days; our gateway enforces a 30-day log retention limit.

What we do not do

• We do not train models on your prompts or completions.
• We do not sell, rent, or share your data with third parties for their own purposes.
• We do not use your content to improve any product, except through the aggregate operational metrics described above.

Caching

For performance we use prefix caching, which keeps recently-seen prompt prefixes in GPU or host memory so repeated context does not need recomputation. This cache is ephemeral, in-memory, and never written to persistent storage. It is destroyed when the serving process restarts.

Infrastructure and sub-processors

Inference runs on GPU servers rented from Vast.ai, in the United States. We prefer verified-datacenter hosts for production traffic. Our API gateway runs on Amazon Web Services in us-west-2 (Oregon, United States). Prompt content transits infrastructure operated by these providers in the course of serving your request.

The inference engine itself is not exposed to the public internet. It is reachable only through our gateway over an encrypted tunnel.

This website

This site sets no cookies and runs no analytics. It is hosted on Cloudflare Pages, which processes visitor IP addresses to serve pages. Pages load web fonts from Google Fonts, which means your IP address is visible to Google when you view this site. No other third party receives data from these pages.

International transfers

Our infrastructure is located in the United States. If you are outside the US, serving your request necessarily involves transferring your prompt to the US for the duration of that request. Because we do not retain content, no personal data remains in the US after the response is returned.

Data protection rights (GDPR / CCPA)

Where GDPR or the CCPA applies, our lawful basis for processing is performance of a contract — serving the inference request you sent. We do not sell personal information as defined by the CCPA, and we do not use it for cross-context behavioural advertising. Because content is not retained, requests to access, port, or delete stored prompt data will generally return nothing; we will respond to the metadata we do hold.

Breach notification

If we become aware of a security breach affecting your data, we will notify affected API customers without undue delay and, where required, the relevant supervisory authority within 72 hours of becoming aware.

Children

The service is not directed to children and is not intended for use by anyone under 18. We do not knowingly collect data from children.

Security

API access requires a bearer credential. Transport is TLS 1.2 or higher. Administrative access to our infrastructure uses AWS Systems Manager; we operate no public SSH.

Your rights

Because we do not retain content, there is generally no stored personal data to access, correct, or delete. For the operational metadata we do hold, contact privacy@neuralaccel.com and we will respond within 30 days.

Changes

We will update the date above when this policy changes. Material changes will be communicated to active API customers.

Contact

privacy@neuralaccel.com — NeuralAccel Corp, California, USA.