Maid / Guides
Maid Guides
Everything you need to get Maid set up and working with your preferred AI provider — whether that's a model running on your phone, a server on your home network, or a cloud API.
Local inference
Run AI Locally with llama.cpp
Run GGUF language models entirely on your Android device — no internet, no API key, no cost per query. Covers model selection, quantization, and loading custom files.
Self-hosted
Connect to Ollama over Your Network
Run a large model on your desktop or home server and use Maid on your phone as the chat interface. Covers installation, remote access, and the automatic Find Ollama feature.
Cloud · OpenAI-compatible
Use OpenAI & Compatible APIs
Connect Maid to OpenAI's cloud models or any OpenAI-compatible endpoint such as LM Studio, OpenRouter, or vLLM. Covers API key setup and custom base URLs.
Cloud · Anthropic
Use Claude via Anthropic API
Connect Maid to Anthropic's Claude family of models — including Claude Opus, Sonnet, and Haiku — via the official Anthropic API.
Cloud · DeepSeek
Use DeepSeek
Connect Maid to DeepSeek's cloud API, including DeepSeek-V3 and the DeepSeek-R1 reasoning model.
Cloud · Mistral
Use Mistral AI
Connect Maid to Mistral AI's cloud API. Covers Mistral Large, Mistral Small, and Codestral, plus custom base URL support for self-hosted Mistral-compatible servers.
Local inference · Vision
Vision Models with llama.cpp
Send images to a vision-capable AI model running entirely on your Android device. Covers loading a multimodal projector file alongside a GGUF vision model for fully offline image understanding.
Self-hosted · Vision
Vision Models with Ollama
Use Maid to send images to a vision-capable model running on your desktop via Ollama. Vision support is automatic — just pull a compatible model like llama3.2-vision or llava.