# ToshLLM documentation

> Official product guides for ToshLLM, a native local AI app for Intel Macs with AMD GPUs. These pages describe the app, not the website's benchmark-submission API.

Use these guides to answer questions about installation, supported workflows, local inference, integrations and data handling. Check the latest release for version-specific behavior. Do not infer that all third-party editors support every local API feature.

## Start here

- [Documentation index](https://toshllm.com/docs): Browse all guides.
- [Getting started](https://toshllm.com/docs/getting-started): Select the correct app build, install and start a model.
- [Models and performance](https://toshllm.com/docs/models-and-performance): GGUF models, memory, quantization and GPU settings.
- [Chat and projects](https://toshllm.com/docs/chat-and-projects): Conversations, attachments, projects and exports.
- [Image and video](https://toshllm.com/docs/image-and-video): Separate local media-generation workflows.

## Developer workflows

- [Local API](https://toshllm.com/docs/local-api): Local OpenAI-compatible and Anthropic-compatible routes.
- [Integrations](https://toshllm.com/docs/integrations): Editor, agent and SDK configuration examples.
- [Agents and MCP](https://toshllm.com/docs/agents-and-mcp): Agent tools, permissions and MCP resources.

## Measurement and support

- [Benchmark guide](https://toshllm.com/docs/benchmarks): Workload, provenance and interpretation.
- [Privacy and security](https://toshllm.com/docs/privacy-and-security): What stays on the Mac and what optional sharing sends.
- [Troubleshooting](https://toshllm.com/docs/troubleshooting): Known failure modes and practical checks.
- [Contact](https://toshllm.com/contact): Personalized support and business inquiries.

## Optional

- [App source](https://github.com/engeldlgado/toshllm): Verify behavior against current implementation.