BlazingNodes is a brand of Spidyhost IT Services Pvt Ltd, a company registered in India. In this policy, “BlazingNodes”, “we” and “us” mean that company.
We run open-weight AI models on our own GPU servers and sell access to them through an API, directly and through marketplaces such as OpenRouter. This policy explains what data we handle, why, and for how long.
The short version
- We do not store your prompts or completions. They are processed in memory to produce a response and are never written to disk, logged, or kept after the request.
- We never use your data to train or improve any model.
- We keep only request metadata, such as time, token counts and API key ID, so we can bill, keep the service reliable and stop abuse. This metadata contains no prompt or completion text.
- Inference runs in the United States (Amazon Web Services, us-east-1).
1. Inference content: prompts and completions
“Inference content” means everything you send to a model (prompts, messages, system prompts, tool definitions, files) and everything the model sends back (completions, tool calls, reasoning text).
- Zero data retention. We process inference content only to generate the response you asked for. We do not write it to disk, databases, logs or backups, and we do not keep it once the request is complete.
- Short-lived memory cache. To answer faster and more cheaply, our serving software may keep parts of recent prompts in GPU memory as a prefix cache. This cache lives only in volatile memory, is never written to persistent storage, and is overwritten as new requests arrive. It is not readable by people and is used only to speed up later requests.
- No training. We never use inference content to train, fine-tune, evaluate or otherwise improve any model.
- No human review. Our staff do not read inference content. Because we do not store it, we cannot retrieve it later, even at your request.
- No end-user identity. We do not need or ask for the identity of the people using an application built on our API. When requests come through a marketplace, we receive them from the marketplace’s servers.
2. Request metadata
For every API request, we record:
- the time of the request;
- a request ID;
- the API key ID (not the key itself);
- the model used;
- the number of input, cached and output tokens;
- the HTTP status;
- timing measurements such as time to first token.
This record contains no prompt or completion text.
We use it to:
- calculate charges and issue invoices;
- measure and improve reliability and speed;
- detect and handle abuse, for example by giving a request ID to a marketplace investigating an account.
We keep this metadata for 24 months, the period we need for billing, disputes and tax records. We then delete it or keep it only in aggregated, anonymous form.
Connection logs. Our servers may briefly record technical connection data for security and debugging: the IP address of the calling system, the time, and the HTTP status, but no content. We keep this for up to 30 days.
3. Customer account and billing information
If you are a direct customer or a marketplace partner, we hold business contact and billing details. That means names, work email addresses, company details, tax identifiers and payment records. We use them to provide the service, invoice you, meet our legal and tax obligations, and contact you about the service. We keep them for as long as you are a customer, and afterwards for as long as company and tax law requires us to keep accounting records.
4. Our website
When you visit blazingnodes.com, our hosting and security provider (Cloudflare) processes technical data such as your IP address to deliver the site and protect it from attacks. We do not use analytics, advertising cookies or third-party scripts, and we do not sell visitor data. Our fonts are served from our own site.
5. Who processes data for us
| Provider | What they do | Data involved | Location |
|---|---|---|---|
| Amazon Web Services | Runs our GPU servers and stores our metadata backups | Inference content (in memory only), request metadata | United States (us-east-1) |
| Cloudflare | Hosts our website and DNS. Cloudflare is not in the path of API requests. | Website visitor technical data | Global network |
| Microsoft (Microsoft 365) | Business email | Customer contact details and correspondence | Microsoft data centres |
We will update this list before adding a new provider, such as a payment processor, and before any new provider handles inference content.
6. Sharing and disclosure
We do not sell personal data, and we do not share inference content with anyone, because we do not keep it. We disclose metadata or account information only:
- to the providers listed above, so they can run the service for us;
- to a marketplace through which you reach us, for billing reconciliation or abuse investigation, limited to request IDs, token counts and timestamps;
- when required by law or a valid legal order;
- to protect the security of the service or other people.
7. International transfers
Inference requests are processed in the United States. If you send requests from another country, your inference content is transferred to the US for the duration of the request only. Request metadata and backups are stored in the US.
8. Security
We protect data with:
- encryption in transit (TLS) on all API and website traffic;
- encrypted storage for metadata and backups;
- API keys stored only as one-way hashes;
- access to production systems limited to authorized staff, with multi-factor authentication.
9. Your rights
Depending on where you live, data protection laws give you rights over your personal data. In India this includes the Digital Personal Data Protection Act, 2023; in the EU and UK, the GDPR. These rights may include access, correction, deletion and objection, and the right to complain to a regulator. To exercise them, contact us using the details below. We hold no inference content, so requests can only cover the account information and metadata we actually keep.
10. Children
Our service is for businesses and developers. It is not directed at children, and we do not knowingly collect personal data from anyone under 18.
11. Changes to this policy
If we change this policy, we will update the date at the top of this page. If a change weakens any commitment in sections 1 or 2, we will tell direct customers and marketplace partners in advance before it takes effect.
12. Contact and grievances
- Privacy questions and requests: privacy@blazingnodes.com
- Grievances: grievance@blazingnodes.com
- Company: Spidyhost IT Services Pvt Ltd, India