I will build a full stack ai web and mobile app with fastapi and langgraph


About this gig
Need a real AI product, not just a demo? I build full-stack AI apps: a web app, an iOS and Android mobile app, and the AI agent backend that powers them.
What I build:
- Web app: Next.js / React with login, dashboard, and a streaming AI chat
- Mobile app: React Native (Expo) for iOS and Android from one codebase
- AI backend: FastAPI + LangGraph/LangChain, or Node.js / NestJS
- AI agent tools on your database or API (PostgreSQL, MongoDB, Supabase)
- Conversation memory, document search (RAG), human approval for writes
- Auth, rate limits, and protected API keys
You receive:
- Full source code you own (web, mobile, backend)
- Acceptance tests with a pass/fail report
- Docker, README, deployment, and TestFlight / Play test builds on Premium
Works with OpenAI, Claude, Gemini, or open-source models.
About me: CTO and full-stack engineer (React, Next.js, React Native, Node.js, NestJS, Python, PostgreSQL, AWS) shipping AI products. See the gallery: a sample agent backend with 18/18 passing tests.
Message me your idea and screens first. I will confirm scope and the right package.
Get to know Muhammad Usman
Full Stack AI Developer for React, FastAPI and AI Agent Apps
- FromPakistan
- Member sinceDec 2024
- Avg. response time1 hour
Languages
Urdu, English
My Portfolio
Other AI Development Services I Offer
FAQ
Which backend do you use?
FastAPI (Python) with LangGraph and LangChain for AI agents, or Node.js / NestJS if your team works in TypeScript. Database: PostgreSQL, MongoDB, Supabase, or Firebase.
Do you build web, mobile, or both?
Both. Web apps in Next.js / React, mobile apps in React Native (Expo) for iOS and Android from one codebase, and the backend they share. Basic covers one platform, Standard a full web or mobile app, Premium web + mobile together.
Which AI models can it use?
OpenAI, Claude, Gemini, Mistral, or open-source models via Ollama, Groq, or OpenRouter. Switching is one setting. Standard and Premium add a fallback model in case your main provider is down.
What do you need from me?
Your idea, a list of screens (a sketch or Figma is great), example AI questions with good answers, and data or API docs. API keys go into your own environment settings, never into chat.
Can the agent change data, or only read it?
Basic is read-only. In Standard and Premium the agent can create or update records through tools we agree on, and each write action waits for a human to approve it in your app before anything is saved.
Where does it run, and what does it cost to run?
On your hosting: Render, Railway, a VPS, AWS, or GCP. You pay the hosting (often $5-$25/month) and model usage directly to those providers. There is no monthly fee to me and no lock-in.
Do I own the code?
Yes. You receive the full Python source code, tests, and Docker setup, and it runs in your accounts with your keys. You can hand it to any developer.
How do you test it?
Each package includes acceptance tests (8, 15, or 25) that run the real API and agent: tool calls, memory, approvals, auth, and errors. You receive the pass/fail report. Premium adds an evaluation set scored against your example answers.
I already have a frontend or an app. Can you add only the AI part?
Yes. I can build just the AI backend with a streaming API for your existing React, Next.js, Flutter, Lovable, or mobile app, or add AI screens to your code. Message me first and I will quote it.
Will you publish my app to the App Store and Google Play?
Premium includes TestFlight and Play internal testing builds and a store submission guide. Submission runs in your own Apple ($99/yr) and Google ($25) developer accounts. Full store submission is available as an extra.

