Operating Model
Different tasks need different things from AI. Some need speed. Some need privacy. Some need the right combination of tools.
SelfReason handles all three from one product — no switching apps, no managing separate subscriptions.
Three ways to work
1. Cloud mode
Use hosted AI when you want the latest models and fast responses.
- Access GPT-4, Gemini, Grok, and more from one interface.
- Low setup, works from any connected device.
- Best for everyday high-volume use.
2. Local mode
Run AI on your Android device when privacy or connectivity matters.
- Data stays on your phone — nothing leaves the device.
- Works offline, on planes, in tunnels, and in restricted networks.
- No subscription required.
3. Orchestrated mode
Route work to the right model, tool, or service automatically.
- Connect to ChatGPT, Gemini, Grok, Codex, Claude Code, and more.
- Switch between models mid-conversation without losing context.
- Use search engines, code assistants, and other tools from one interface.
Why these modes work together
Most products force you into one mode. SelfReason lets you move between them based on what the moment needs.
- Speed matters? Use cloud mode.
- Privacy matters? Switch to local.
- Task spans multiple tools? Use orchestration.
You do not need to pick one approach. SelfReason adapts to how you work.
Related pages
- Read Product Overview for the full product story.
- Read Agent Orchestration for how multi-model routing works.
- Read Offline AI on Android for the local mode experience.