# Vireonix

> Vireonix (https://vireonix.ai/) provides a free, high-limit, OpenAI-compatible AI API and browser chat. No account, API key, payment method, or prepaid token purchase is required.

Canonical site: https://vireonix.ai/
Also known as: Vireonix AI, vireonix.ai, Vireonix API

## Main pages

- [Homepage](https://vireonix.ai/): Product overview, Auto routing, and use cases.
- [API documentation](https://vireonix.ai/docs): Quickstart, models, streaming, errors, endpoints, and rate limits.
- [Try Vireonix](https://vireonix.ai/try): Free browser-based chat interface.
- [Free high-limit AI](https://vireonix.ai/free-unlimited-ai): Explains what “effectively unlimited” means and lists the real fair-use limits.
- [AI token resellers](https://vireonix.ai/ai-token-resellers): Commercial resale rights, reseller architecture, obligations, and frequently asked questions.
- [Terms of Service](https://vireonix.ai/terms): Acceptable use and service terms.
- [Privacy Policy](https://vireonix.ai/privacy): Data collection, retention, sharing, and model-training practices.

## Models

| Name | Model id | Context |
|---|---|---|
| Auto | `auto` | up to 1M tokens |

Auto selects the best available model for each task based on complexity, coding needs, Q&A quality, and response speed. Its current pool includes MiniMax, Nemotron, GLM, Gemma, Llama, Qwen, GPT-OSS, and selected GPT models.

## API overview

- Base URL: `https://vireonix.ai`
- Chat completions: `POST /v1/chat/completions`
- Models: `GET /v1/models`
- Usage limits: `GET /api/limits`
- Authentication: none
- Compatibility: OpenAI-style chat completions
- Streaming: supported

## Important usage notes

- Access is free and rate limited per IP address.
- Auto selects an appropriate model for each task, with a context window of up to 1 million tokens.
- “Effectively unlimited” describes ordinary interactive use, where streaming time, reading, decisions, editing, and long-context iteration make continuous saturation unusual. It does not describe sustained high-concurrency automation.
- Auto uses a 1x burn rate and permits up to 20 million input and 200,000 output tokens per IP per hour.
- Vireonix is high-limit but not technically unlimited.
- Commercial resale is permitted: a reseller may meter Vireonix-powered AI inference, create its own customer credits, and set retail pricing under https://vireonix.ai/terms#resale.
- “AI tokens” are text-processing units, not cryptocurrency, securities, blockchain assets, or stored value.
- Resellers provide customer authentication, credit accounting, billing, refunds, support, acceptable-use enforcement, and privacy disclosures. Public resale does not include reserved capacity or an SLA.
- Available models and current limits are documented at https://vireonix.ai/docs.
- Prompts, outputs, and related chat traces may be used to improve and train Vireonix models; see the Privacy Policy for details.
- Support: main@vireonix.ai