Xlance Collective · Client Project · Feb–Apr 2024
TalkTask
Create a task manager that reduces the friction of entering and organizing work by allowing users to speak naturally, turn one recording into structured tasks, and continue capturing ideas even when connectivity is unavailable.
Role
Sole App Developer (Designer-Led Screens)
Timeline
Feb – Apr 2024
Scope
Mobile (Flutter)
Status
Handed Over — Closed Testing
Challenge
Create a task manager that reduces the friction of entering and organizing work by allowing users to speak naturally, turn one recording into structured tasks, and continue capturing ideas even when connectivity is unavailable.
Role & Collaboration
I was the sole application and feature developer, working from screens created by a UI/UX designer — the same designer I collaborated with on AMZ Whisper. A teammate provided limited platform and release-configuration support. This was the final project I handled at Xlance Collective, handed over as an almost-finished, closed-testing build before I left.
Solution
I built a local-first task system with categories, importance, due dates, search, completion, soft deletion, restoration, and calendar views. I then connected audio recording, speech transcription, and structured AI processing so one spoken note could produce multiple editable tasks. For unreliable connectivity, recordings could be cached locally and processed later once the device came back online.
Technical Contribution
The application also included social authentication, Firebase-backed profiles and profile images, device-calendar synchronization, subscription tiers with purchase restoration, onboarding, and English and Chinese localization. I translated the designer’s screens into reusable Flutter components and refined the product through its closed-testing preparation.
- Voice-to-task workflow: recording → transcription → AI-structured multi-task output → review → local storage
- Offline audio caching with deferred processing once connectivity returns
- Local-first SQLite task and category architecture with full lifecycle (complete, soft-delete, restore)
- Device-calendar synchronization and social sign-in (Google, Facebook, Apple)
Outcome
The core personal task-management and voice-to-task experience was substantially implemented and handed over in an almost-finished, closed-testing state. Planned team-collaboration and notification features remained hidden or prototype-level at handover — an honest boundary worth stating plainly rather than overselling a fully finished product.
Voice-to-task workflow spanning recording, transcription, AI structuring, review, and storage.
Tech Stack
Mobile
Voice / AI
Backend