AI English Learning App for Developers
Budget: $30 – $250 USD
(Document Attached)
Objective
We are developing a mobile app similar to Loora, focused on teaching English to software developers through
interactive AI-driven conversations based on real IT scenarios.
Tech Stack
Frontend: Flutter (iOS & Android)
Backend-as-a-Service: Supabase (PostgreSQL, Auth, Realtime, Storage)
AI Services: OpenAI GPT-4 (text), Whisper (speech-to-text), ElevenLabs or Azure (text-to-speech), D-ID API for avatar
video.
Core Features (MVP Phase)
1. Authentication & User Profile
2. AI Chatbot with predefined topics
3. Voice interaction (STT + TTS)
4. Topic and level-based learning
5. Progress Reports
6. Talking Avatar Video Integration using D-ID
Talking Avatar Feature (Priority in MVP)
Goal: Enable a talking AI video avatar that reads GPT-4 replies with facial expressions.
Video generated by D-ID API from text + TTS audio.
Flutter frontend plays the video in a chat container.
Show loading while rendering video.
Fallback to voice-only if video fails.
Ensure responsiveness, mute/unmute control, and local caching.
UI/UX Expectations
Modern, clean design (Figma will be provided)
Responsive chat interface
Smooth animations and transitions
UX must support AI feedback and user correction.
Functional Requirements
Offline history
Topic-based progress tracking
Real-time interaction
Support for switching between text and voice
Milestone-based integration tests
Non-Functional Requirements
Clean, documented codebase
Secure API key handling
Use of riverpod or bloc architecture
Git version control with commit history
Unit and widget testing required
Deliverables
Each milestone must include:
- Working build (APK/TestFlight)
- Source code access
- Release notes
- Demo session
- Developer QA test report
Milestones
M1 (Week 1-2): Auth & Profile Setup
M2 (Week 3-4): Chat + GPT Integration + Avatar Video
M3 (Week 5-6): Voice Features + Topic Engine
M4 (Week 7-8): Progress Reports + Final QA
M5 (Optional): Admin Panel
Expectations from Freelancer
- High-quality Flutter code
- Weekly or async video demos
- Quick responses to feedback
- Clean and scalable architecture
- Active communication
To Include in Your Proposal
- Flutter portfolio (AI/chat apps preferred)
- GitHub/code samples
- Experience with Supabase, GPT-4, D-ID
- Development methodology and tools
- Timeline breakdown with milestone estimates
AI Optimization and Cost-Saving Strategy
To reduce unnecessary API calls and optimize costs, the system must store and reuse AI-generated content whenever
possible. This includes:
- Caching responses to repeated questions or common prompts.
- Sharing AI-generated lessons and dialogues across users with similar levels.
- Creating a local knowledge base of AI-generated corrections, topic templates, and feedback.
- Using metadata tags to classify and retrieve cached answers efficiently.
This approach ensures faster responses for users and significantly reduces spending on GPT, TTS, and avatar
Objective
We are developing a mobile app similar to Loora, focused on teaching English to software developers through
interactive AI-driven conversations based on real IT scenarios.
Tech Stack
Frontend: Flutter (iOS & Android)
Backend-as-a-Service: Supabase (PostgreSQL, Auth, Realtime, Storage)
AI Services: OpenAI GPT-4 (text), Whisper (speech-to-text), ElevenLabs or Azure (text-to-speech), D-ID API for avatar
video.
Core Features (MVP Phase)
1. Authentication & User Profile
2. AI Chatbot with predefined topics
3. Voice interaction (STT + TTS)
4. Topic and level-based learning
5. Progress Reports
6. Talking Avatar Video Integration using D-ID
Talking Avatar Feature (Priority in MVP)
Goal: Enable a talking AI video avatar that reads GPT-4 replies with facial expressions.
Video generated by D-ID API from text + TTS audio.
Flutter frontend plays the video in a chat container.
Show loading while rendering video.
Fallback to voice-only if video fails.
Ensure responsiveness, mute/unmute control, and local caching.
UI/UX Expectations
Modern, clean design (Figma will be provided)
Responsive chat interface
Smooth animations and transitions
UX must support AI feedback and user correction.
Functional Requirements
Offline history
Topic-based progress tracking
Real-time interaction
Support for switching between text and voice
Milestone-based integration tests
Non-Functional Requirements
Clean, documented codebase
Secure API key handling
Use of riverpod or bloc architecture
Git version control with commit history
Unit and widget testing required
Deliverables
Each milestone must include:
- Working build (APK/TestFlight)
- Source code access
- Release notes
- Demo session
- Developer QA test report
Milestones
M1 (Week 1-2): Auth & Profile Setup
M2 (Week 3-4): Chat + GPT Integration + Avatar Video
M3 (Week 5-6): Voice Features + Topic Engine
M4 (Week 7-8): Progress Reports + Final QA
M5 (Optional): Admin Panel
Expectations from Freelancer
- High-quality Flutter code
- Weekly or async video demos
- Quick responses to feedback
- Clean and scalable architecture
- Active communication
To Include in Your Proposal
- Flutter portfolio (AI/chat apps preferred)
- GitHub/code samples
- Experience with Supabase, GPT-4, D-ID
- Development methodology and tools
- Timeline breakdown with milestone estimates
AI Optimization and Cost-Saving Strategy
To reduce unnecessary API calls and optimize costs, the system must store and reuse AI-generated content whenever
possible. This includes:
- Caching responses to repeated questions or common prompts.
- Sharing AI-generated lessons and dialogues across users with similar levels.
- Creating a local knowledge base of AI-generated corrections, topic templates, and feedback.
- Using metadata tags to classify and retrieve cached answers efficiently.
This approach ensures faster responses for users and significantly reduces spending on GPT, TTS, and avatar