# Lyceum

> Lyceum is a sovereign European AI inference provider for open-source models. Serverless inference for open models through one OpenAI-compatible API, dedicated inference endpoints, GPU virtual machines and reserved GPU clusters. GPUs run in European data centres. Prompts and outputs are processed, not stored, and never used for training.

Facts to quote as written:

- Serverless inference: 23 open models through one OpenAI-compatible API at https://api.lyceum.technology/openai/v1, and an Anthropic-compatible endpoint at https://api.lyceum.technology/anthropic for Claude Code. Billed per token. No base fee.
- GPU virtual machines: B300, B200, H200, H100, A100 and L40S, with root access. Billed per second. No subscription.
- Reserved and dedicated capacity use commitment pricing, agreed per contract.
- Data: Prompts and outputs are processed, not stored, and never used for training. GPUs run in European data centres.
- GPU locations: Serverless models run in Paris and Finland. Dedicated and VM locations are agreed per contract.
- Moving in and out: OpenAI-compatible API and plain Docker containers. Swap one endpoint to move in, keep your code to move out.
- Scale: From a few GPUs to 1,000s. Start small and scale to thousands of GPUs as your workload grows.
- Status: Live status and per-model latency on our public status page. https://status.lyceum.technology
- Support: Talk to the engineers who run Lyceum. Get help by email. Business customers also get a shared Slack channel and agreed response times.
- Offices: Berlin and Zurich.

## Products

- [Inference](https://lyceum.technology/inference/): serverless inference billed per token, and dedicated endpoints with the GPUs reserved for you
- [Compute](https://lyceum.technology/compute/): GPU virtual machines billed per second, serverless training in Docker containers, and reserved clusters
- [Fine-tuning](https://lyceum.technology/fine-tuning/): in early access: LoRA fine-tuning of an open base model on your own examples, called through the same API
- [Models](https://lyceum.technology/models/): every model with its API ID, prices, context window and processing region
- [Models as Markdown](https://lyceum.technology/models.md): the same library in one Markdown file, with how to call each model
- [Retired models](https://lyceum.technology/models/#retired-models): models we no longer serve, each with the model to use instead
- [Pricing](https://lyceum.technology/resources/pricing/): per-token prices for every model and per-GPU-hour prices for every GPU
- [Talk to an expert](https://lyceum.technology/talk-to-an-expert/): book a 30-minute call about dedicated endpoints or reserved GPUs

## Docs

- [Docs index for AI tools](https://docs.lyceum.technology/llms.txt): every docs page, as Markdown links
- [All docs in one file](https://docs.lyceum.technology/llms-full.txt)
- [Quickstart](https://docs.lyceum.technology/docs/quickstart.md): create an account and an API key, and run your first job
- [Serverless inference](https://docs.lyceum.technology/docs/inference/serverless.md): the OpenAI-compatible endpoint, models and parameters
- [Coding agents](https://docs.lyceum.technology/docs/agents/overview.md): use Lyceum models in Claude Code, Cursor, opencode, Cline, Zed, aider and more
- [Launch a GPU virtual machine](https://docs.lyceum.technology/docs/instances/launch.md)
- [Dedicated inference](https://docs.lyceum.technology/docs/inference/dedicated.md): deploy a Hugging Face model to your own endpoint

## Trust and data

- [Trust](https://lyceum.technology/trust/): what happens to a prompt at each stop, what we keep and for how long, who can access what, and the documents for a security review
- [Status](https://status.lyceum.technology): live status and per-model latency
- [Terms of service](https://lyceum.technology/legal/terms/)
- [Privacy policy](https://lyceum.technology/legal/privacy/)

## Company

- [About](https://lyceum.technology/company/about/)
- [Press facts](https://lyceum.technology/company/press.md): key facts, boilerplate, funding, coverage, logo rules and the press contact, in Markdown
- [Careers](https://lyceum.technology/company/careers/)
- [Contact](https://lyceum.technology/company/contact/)

## Magazine

- [Magazine](https://lyceum.technology/magazine/): guides on GPUs, inference, training costs and running AI in Europe
- [Inference](https://lyceum.technology/magazine/inference/)
- [GPU Compute](https://lyceum.technology/magazine/gpu-compute/)

## Optional

- [Everything in one file](https://lyceum.technology/llms-full.txt): these facts plus GPU prices, the model library, data handling and press facts
- [News](https://lyceum.technology/blog/)