# Backend Installation

## Ollama

Ollama is the default backend installed by the platform builders.

```text
backends/ollama/windows/ollama.exe
backends/ollama/macos/ollama
backends/ollama/linux/ollama
models/ollama/
```

Launchers set `OLLAMA_MODELS` so models remain inside the JackAILocal target.

## llama.cpp

The Rust runtime can call a reachable OpenAI-compatible `llama.cpp` server at `LLAMA_CPP_URL`. This is an optional local chat fallback and is not required by the standard builder payload.

## Voice

```text
backends/whisper.cpp/<platform>/whisper-cli
backends/piper/<platform>/piper
models/whisper/ggml-base.bin
models/piper/en_US-libritts_r-medium.onnx
```

`windows/Install-Voice.ps1` and `unix/Install-Voice.sh` download verified official assets. Linux and macOS build whisper.cpp from the pinned source tag before installing the model assets.