One container. The whole stack.
Basilisk is a complete local LLM server and analytics engine, packaged as a single Docker container.
Drop it on any machine — laptop, server or edge device — and you have OpenAI-compatible chat and completion endpoints, plus time-series database analysis and anomaly scoring. No cloud, no per-token fees, no data leaving your network. Private intelligence, where you need it.
Chat & completions
A drop-in OpenAI-compatible API. Use it with LangChain, LlamaIndex, your own scripts or any existing OpenAI client — just change the base URL.
Time-series analysis
Native integration with VictoriaMetrics, Prometheus and other time-series databases. Ask natural-language questions about your metrics and get real answers.
Anomaly scoring
Anomaly detection on your time-series data, using the LLM together with statistical models — scored alerts with plain-English explanations.
Simple to run
One command. Works on x86-64 and ARM — laptops, servers, Kubernetes, even a Raspberry Pi.
Private by design
Your data never leaves your infrastructure. Suited to sensitive environments, regulated industries, and anyone tired of cloud bills.
Model-flexible soon
Works with any GGUF model — Llama 3, Mistral, Phi, Gemma and more. Swap models by changing one environment variable.
Cloud LLMs are expensive. Basilisk is not.
Run capable local inference with full control over cost, latency and data sovereignty.
Ready to run your own?
Basilisk is in private beta. If you would like early access, a demo, or to run it in your own environment — let's talk.