docs/doc/developer/apps/examples/OmiMentor.mdx
The app intelligently monitors and analyzes your conversations:
You: "I'm not sure if I should take this new job offer..."
[AI listens and analyzes conversation]
Friend: "What's holding you back?"
[Continues collecting context]
You: "The salary is better but I'd lose my work-life balance..."
[AI evaluates if intervention is needed]
AI Mentor Decision:
AI Sends: "You just said work-life balance matters more than money. The answer's already clear - you're trying to convince yourself otherwise. What would you tell your best friend in this situation?"
→ Notification sent! </Note>
The app uses OpenAI for intelligent mentorship:
Interruption Criteria:
<CodeGroup> ```text Requirements (ALL must be met) 1. User is actively participating in the conversation 2. User expressed a specific problem, challenge, goal, or question 3. AI has a STRONG, CLEAR opinion that would significantly impact the situation 4. The insight is time-sensitive and worth interrupting for ```- Speaks DIRECTLY to the user (no third-person analysis)
- Takes a clear stance (no "however" or "on the other hand")
- Under 300 characters
- Simple, everyday language
- References specific details from what user said
- Bold and direct advice
- Ends with a specific implementation question
OMI Mentor uses an intelligent message buffering system:
Smart Collection:
The AI uses multiple sources to personalize advice:
Topics are automatically extracted and used to: