Custom Connection Guide

Use a custom connection when your provider is not listed directly, or when you run a server that exposes an OpenAI-compatible or Anthropic-compatible API.

What you need

• A base URL, such as https://api.example.com or http://192.168.1.20:5000
• The correct route prefix, such as /v1/, /api/, or /api/v1/openai/
• The matching API profile, usually OpenAI or Anthropic
• An API token if the server requires authentication

Add the connection

1. Open 3sparks Chat and go to Connections.
2. Tap + and choose Other (Custom).
3. Leave the name blank for a generated name, or enter your own.
4. Enter the server host in Base URL.
5. Choose the correct API Profile.
6. Enter the Route Prefix used before endpoints like models or chat/completions.
7. Add an API Token if required. Tokens are stored in Keychain.
8. Tap Create Connection.
9. Use Fetch Models when the server supports a models endpoint, then choose a default model.

Common route examples

• OpenAI-compatible: base URL api.example.com, route prefix /v1/
• OpenWebUI: base URL YOUR_SERVER:8080, route prefix /api/
• AnythingLLM: use the OpenAI-compatible endpoint shown in AnythingLLM settings

Troubleshooting

• Phones cannot reach a desktop server at localhost; use a LAN IP or Tailscale IP.
• If model fetch fails but chat works, enter a model manually or choose one from the provider dashboard.
• If requests fail immediately, verify HTTPS settings, route prefix, and API profile.
• If the model warns about context size, lower the token limit for the connection.