BirthdayAI
An AI-powered birthday companion that stores context about the people in your life and drafts hyper-personal messages using that context, so you never send a generic 'Happy Birthday' again.
Overview
BirthdayAI is a birthday companion app that closes the gap between getting a notification and actually sending something personal. It stores a profile for each person in your life: their interests, inside jokes, things they hate, memories you share, and when their birthday approaches, surfaces that context alongside an AI-drafted message that sounds like you wrote it at your thoughtful best, not you half-awake at 9am.
Why it was built
The problem isn't forgetting birthdays. Google Calendar handles that. The problem is the 90-second decision paralysis between the notification and actually sending something meaningful. This app is built to collapse that gap entirely, designed for relationship-conscious professionals who have long contact lists and short morning windows.
Contact Context Profiles
Each contact has a profile card where you log their interests, shared memories, what they care about, and relationship context. The profile is the memory layer that makes the AI output actually personal.
Smart Birthday Surfacing
Integrates with Google Calendar to pull birthday events. Surfaces reminders 3 days in advance with the person's profile pre-loaded, not a ping on the day, but enough lead time to do something meaningful.
AI Message Drafting
Feeds the contact's profile to Gemini with a prompt tuned for tone: warm but not over-the-top, personal not generic. Generates 3 message variants (short, medium, heartfelt) that can be edited before sending.
Tech Breakdown
Reads the user's existing birthday calendar entries rather than asking them to re-enter dates. OAuth integration means the app stays in sync automatically as contacts are updated.
The message drafting prompt is the core IP of the app. Gemini was chosen for its long context window: the full contact profile can be passed as context without truncation, which directly improves message quality.
Contact profiles and message history need server-side persistence with per-user isolation. PostgreSQL handles relational data (users, contacts, birthdays, messages) cleanly with a Prisma ORM layer.
For sending directly from the app rather than deep-linking to the user's own WhatsApp. Pending API access approval, so the share-link approach is the MVP fallback.
- 1
The quality of AI-drafted messages is entirely dependent on the quality of context stored. The onboarding flow has to convince users to invest 2-3 minutes per contact profile upfront. Without that, the output is only marginally better than a generic message.
- 2
Google Calendar OAuth requires handling token refresh, scope re-consent on permission changes, and graceful degradation when the user hasn't granted calendar access. All three are planned for the auth flow.
Currently in PRD and planning phase. Phase 1 covers contact profiles, birthday surfacing from Google Calendar, and AI message drafting. Phase 2 adds the WhatsApp send integration and a relationship strength score based on how often each contact has been messaged through the app.
Want to know more?
Happy to walk you through the details, just reach out.