Xyntetik

Local models on hardware you own. Xyntetik is independent and bootstrapped: the engine is free forever under Apache 2.0, and consulting and enterprise work fund the hardware. Built in Sweden. Your data never leaves the building.

Runner

A single-binary GGUF inference and LoRA training engine written from scratch in plain C, for CPU, CUDA and Metal. It trains LoRA adapters through the same quantized weights it serves, closes tool calls that run out of tokens so they still parse, and records every run as a receipt that replays. Every claim on these pages is tied to a measurement you can re-run.

• Runner: what it is, sixty seconds to a served model, why this and not llama.cpp
• Train a LoRA on the quantized GGUF you serve
• Reproducible LoRA training with receipts
• Tool calls that survive the token limit
• Benchmarks, models and published artifacts, support matrix, all docs
• Source, releases and issues on GitHub

Research

Measured reports and artifacts are published on Hugging Face: quantization fidelity ladders, pruned and surgically reduced models with their envelopes, and the adapters the training claims are measured on.