Getting Started

3sparks Chat can use on-device models, local servers, and cloud AI providers. The setup path depends on where the model runs.

1. Choose how you want to run AI

• On device: use Local Models to download and load supported GGUF models directly on your phone.
• Local server: use LM Studio, Ollama, or another server on your computer.
• Cloud API: use OpenAI, Anthropic, Mistral, Gemini, OpenRouter, Azure OpenAI, or a custom compatible endpoint.

2. Set up the model

For on-device models:

1. Open Local Models.
2. Browse Available models or use the filter field.
3. Download a supported model.
4. Open Downloaded and tap Load.
5. Start chatting with the loaded local model.

For local-server or cloud providers:

1. Open Connections.
2. Tap +.
3. Choose a provider.
4. Enter the base URL or API key the provider needs.
5. Leave the name blank to let 3sparks Chat generate one, or enter your own.
6. Tap Create Connection.
7. Tap Fetch Models and choose a default model.
8. Tap Done.

API tokens are stored securely in Keychain. Provider-specific setup details live in the Connections guide and the provider articles.

3. Start a chat

1. Open the chat list.
2. Start a new chat.
3. Pick the loaded Local Model, local-server connection, or cloud connection you want to use.
4. Type a message and send it.

You can add files, images, PDFs, web pages, and location from the + menu in the composer.

4. Tune the chat

Open chat settings to change the chat name, instruction, model or provider, streaming, temperature, output tokens, reasoning, context length, or Local Knowledge settings when those controls are available.