Intelligent AI Writing Assistant

Problem

Users needed AI assistance without leaving their typing workflow.

Solution

Integrated LLM APIs directly into a native keyboard extension supported by scalable backend services.

Solution Architecture

Keyboard extension communicates with a lightweight backend that proxies LLM requests. Caching and debouncing minimize latency while keeping token usage efficient.

Mobile + Backend Design

Native keyboard extension with secure keychain storage for API credentials. Backend handles rate limiting and context window management.

AI / Python Services

Optional Python service for prompt engineering and response post-processing before delivery to the client.

Technical Challenges

Balancing response time with context length, and maintaining a smooth UX when the network is slow.

Outcome

Low-latency AI typing experience with scalable architecture.

Key features & tech

• Real-time AI suggestions
• Context-aware generation

• SwiftUI
• Firebase
• LLM APIs