Question 1

What is Parta Runtime Lite?

Accepted Answer

Parta Runtime Lite is a single, portable executable that runs a local AI inference server on your own machine. It is a stripped-down version of Parta Runtime, the inference engine behind the Parta platform, with the identity layer, audit chain and policy engine removed. There is no installer and no background service: copy the executable anywhere, point it at a model file, and run it.

Question 2

Which API formats does it support?

Accepted Answer

Parta Runtime Lite speaks OpenAI-compatible and Anthropic-compatible chat completion formats, so tools and SDKs already built against those APIs work against a local Parta Runtime Lite server without modification. Vision support is included for models that accept image inputs.

Question 3

Does it include the identity, audit or policy features of Parta?

Accepted Answer

No. Parta Runtime Lite is intentionally stripped of the identity layer, audit chain and policy engine that make up production Parta Runtime. It is a plain inference server only, built for people who want fast local inference without any of the platform infrastructure.

Question 4

What hardware does it require?

Accepted Answer

The first release is a CUDA build for Windows, requiring an NVIDIA GPU with a working driver. Vulkan and Linux builds are on the roadmap for hardware and platforms outside that combination.

Question 5

How do I run it?

Accepted Answer

Download the zip, extract it anywhere on disk, point the included configuration file at a GGUF model file you already have, and run the executable. There is no setup beyond that: no installer, no admin rights, no background service.

Parta Runtime Lite

Parta Runtime,
without the platform.

Speaks the formats
you already use.

Download, unzip,
run.

CUDA is first.
Vulkan and Linux next.

Download and verify.

Parta Runtime Lite

Parta Runtime,without the platform.

Speaks the formatsyou already use.

Download, unzip,run.

CUDA is first.Vulkan and Linux next.

Download and verify.

Parta Runtime,
without the platform.

Speaks the formats
you already use.

Download, unzip,
run.

CUDA is first.
Vulkan and Linux next.