Last updated: September 20, 2026
These terms are a practical operating policy for using Token Factory, not legal advice.
Token Factory is an experimental research service providing access to self-hosted large language model inference on our own GPU cluster. Available models, limits, features, latency, throughput, output quality, and routing behavior may change without notice; there is no performance guarantee.
You must be at least 18 years old to create an account or use the service. Beta access may be limited or revoked at our discretion.
You agree not to use the service for unlawful purposes, not to attempt to disrupt the infrastructure, not to resell access without written permission, and not to use it to generate content that violates applicable law.
The service runs on research infrastructure (GPU allocations that periodically expire). Outages, model swaps, and capacity changes will happen. We publish live status on the status page. No SLA is offered during beta.
We store request metadata (timestamps, token counts, latency, error rates) for operations. We do not use your prompts or completions for training. Do not send secrets, credentials, or regulated data through the service.
Model outputs are generated by statistical systems and may be incorrect, biased, or nonsensical. Verify anything important. You are responsible for how you use the outputs.
We may update these terms; material changes will be noted on this page with a new date. Continued use after changes means acceptance.