I am building an MVP chatbot app designed for people to ask questions and receive helpful, conversational answers. The chatbot will use LLM APIs (Grok and Claude) to generate responses. The goal is to create a simple but reliable chat interface where users can type in their queries, get AI-generated replies, and review their chat history securely.
The focus of this MVP is clarity, functionality, and clean UI — nothing fancy or overcomplicated, just a solid base that looks neat and works well. Once this base version is ready, I will handle the detailed design and future customizations.
Scope of Work
UI / Frontend (FlutterFlow)
Onboarding & Login (basic auth with Firebase/email/phone). Consent Screen (user agrees to Privacy & Data Use). Main Chat Screen User input + chatbot response bubbles. Typing indicator + streaming responses (if possible). Feedback Option Thumbs up/down per response. Minimal, clean UI (no heavy design, but should not look broken).
Backend / Database
Firebase/Firestore (preferred) or cheapest alternative (AWS/Supabase). Store: users → userID + consent timestamp. messages → linked to userID, with 1:1 mapping of each prompt and response stored in separate fields under the same record (prompt, response, timestamp). Privacy: Users can only see their own chats. Option to delete account + data. API Key Security: Keys stored securely in backend (Firebase Functions or equivalent).
AI Integration
Integrate Grok API (default for testing). Integrate Claude Sonnet 4.0 API (as fallback/higher level). Admin-only toggle to switch between Grok and Claude (end users should not see this option). Context Handling: Last 10–15 messages included in conversation so chatbot maintains flow. Session Summary (Optional): On app reopen, display a summary of last 10–15 chats.
Deliverables
Full FlutterFlow project (transferred to my account). Backend setup (Firebase or cheapest recommended). Working chatbot with both Grok & Claude APIs integrated. Test-ready app sharable with ~10–15 users. Documentation: Setup guide (backend + API keys). How to switch models (admin-only). Loom/video walkthrough.
Timeline Expected delivery: 2–4 days for the MVP version. Additional time (up to 5–7 days) may be required if extra polish or optional features (like streaming responses or session summaries) are included.
Corporate Human Capital Metrics Deck Category: Business Analysis, Corporate Identity, Data Visualization, Excel, Google Slides, Infographics, Powerpoint, Research Writing Budget: $200 - $250 SGD
Recreational Lake Permit Plan Category: 3D Modelling, 3D Rendering, AutoCAD, Building Architecture, Civil Engineering, Environmental Engineering, Landscape Design, SketchUp Budget: $250 - $750 USD
1 Minute Fundraising Video Creation Category: After Effects, Animation, Script Writing, Video Ads, Video Editing, Video Post Editing, Video Production, Visual Effects Budget: $10 - $20 CAD
