Features
Multi-Provider Chat
Sythoria unifies multiple AI providers under a single interface. Switch between OpenAI, Anthropic, Gemini, Ollama, and more — without switching tabs, apps, or accounts.
Switching providers
Use the model dropdown in the chat input bar to switch models at any time — even mid-conversation. Each new message uses the currently selected model, so you can:
- Start a conversation with GPT-4o, then switch to Claude for a second opinion
- Use Ollama for offline work, then jump to Gemini for a complex query
- Compare outputs from different models by asking the same question with different providers
Switching is instant. There's no loading, no reconnection — just select and send.
Supported providers
| Provider | API format | Local/Cloud | API key required |
|---|---|---|---|
| OpenAI | OpenAI API | Cloud | Yes |
| Anthropic | Anthropic API | Cloud | Yes |
| Google Gemini | OpenAI-compatible | Cloud | Yes |
| Ollama | OpenAI-compatible | Local | No |
| OpenRouter | OpenAI API | Cloud | Yes |
| NVIDIA NIM | OpenAI API | Cloud | Yes |
| Custom | OpenAI-compatible | Any | Optional |
Adding multiple models
You can add as many model configurations as you want in Settings:
- Open Settings from the sidebar
- Click Add Model
- Choose a Provider Preset or set up a custom endpoint
- Enter the model details and API key
- Repeat for each model you want to use
Each model gets its own connection indicator (green = connected, red = error, yellow = connecting).
Custom providers
Sythoria supports any OpenAI-compatible API. Select Custom from the provider dropdown and enter:
| Field | What to enter | Example |
|---|---|---|
| API Base URL | The endpoint URL | https://my-llm.example.com/v1/chat/completions |
| API Key | Your key (if required) | sk-... |
| Model ID | The model identifier | my-model-v2 |
This works with services like:
- Together AI —
https://api.together.xyz/v1/chat/completions - Groq —
https://api.groq.com/openai/v1/chat/completions - Fireworks AI —
https://api.fireworks.ai/inference/v1/chat/completions - vLLM — your own server endpoint
- LM Studio —
http://localhost:1234/v1/chat/completions - Any other OpenAI-compatible server
