One familiar API
Use OpenAI-compatible chat completions and switch models by changing one model ID.
Models
Frontier models, run on the Lilac network.
Search models available through Lilac. Compare context, capabilities, and per-token pricing.
Loading the live model catalog…
Fetching the models currently available through Lilac.
One API
Switch models without switching integrations
Build against a consistent interface while choosing the context, capabilities, and price that fit each workload.
Zero data retention
Prompt and completion data is never stored or used to train models.
Transparent token pricing
Compare input, cached-input, and output prices before you send a request.
Inference API
Choose a model and put idle GPUs to work.
Create an API key, copy a model ID, and send your first OpenAI-compatible request.