You will get Custom AI Agent & MCP Integration for Your App (RAG · Next.js · Python)
Rising Talent
Project details
I add AI that actually does the work not a toy chatbot.
I build production AI features into the app you already run:
• Agents that plan, search your data, and take multi-step actions
• RAG assistants grounded in your docs/site, with citations (no hallucinations)
• Custom MCP servers that connect Claude Desktop, IDEs, or your agents to your tools and data
• Streaming, error handling, deployed live
Proof, not promises: a live agent demo (DocChat Agent) and an open-source MCP server are in my portfolio links on request.
Stack: Next.js, TypeScript, Python/FastAPI, Vercel AI SDK, OpenAI / Gemini / Claude. You keep your own API keys.
Message me with what your app does and what you want the AI to handle I'll tell you exactly how I'd build it.
I build production AI features into the app you already run:
• Agents that plan, search your data, and take multi-step actions
• RAG assistants grounded in your docs/site, with citations (no hallucinations)
• Custom MCP servers that connect Claude Desktop, IDEs, or your agents to your tools and data
• Streaming, error handling, deployed live
Proof, not promises: a live agent demo (DocChat Agent) and an open-source MCP server are in my portfolio links on request.
Stack: Next.js, TypeScript, Python/FastAPI, Vercel AI SDK, OpenAI / Gemini / Claude. You keep your own API keys.
Message me with what your app does and what you want the AI to handle I'll tell you exactly how I'd build it.
Programming Languages
JavaScript, Python, TypeScriptCoding Expertise
Localization, Performance Optimization, SecurityWhat's included
| Service Tiers |
Starter
$150
|
Standard
$650
|
Advanced
$1,500
|
|---|---|---|---|
| Delivery Time | 4 days | 9 days | 16 days |
Number of Revisions | 1 | 2 | 3 |
Number of Pages | 1 | 1 | 1 |
Design Customization | - | - | |
Content Upload | - | - | - |
Responsive Design | - | ||
Source Code |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$50 - $250
Additional Revision
+$30
Design Customization
(+ 3 Days)
+$100
Extra data source / integration
(+ 3 Days)
+$120
Add an agent action (tool)
(+ 3 Days)
+$150
Custom MCP server
(+ 5 Days)
+$350Frequently asked questions
About Phisit
AI Developer & LLM Security | RAG, Agents, Python/Next.js
Mueang Nonthaburi, Thailand - 8:24 am local time
Most developers can wire up a chatbot. Few can also tell you how it leaks
data, obeys a hidden instruction buried in a document, or hands one user
another user's answer. I do both.
WHAT I BUILD (Python / FastAPI and TypeScript / Next.js):
- LedgerLens - an AI bookkeeping tool with a self-audit engine that flags
fraud and anomalies. Deterministic pandas core + Claude, 33 automated tests.
- DocChat - a RAG agent that answers questions about any webpage with
grounded, cited answers.
- INKVERSE - a full content platform I built and launched solo in ~2 weeks:
authentication, PromptPay payments, an Android app, and an anti-scraping
reader. I then drove 100,000+ views to it.
HOW I SECURE AI APPS:
I red-team LLM applications against the OWASP LLM Top 10 - prompt injection
(direct and indirect), data leakage, and jailbreaks - and I write the fixes,
not just the report. Open-source tools I built:
- promptraid - an LLM prompt-injection red-team toolkit with canary-confirmed
probes across the OWASP LLM Top 10.
- raidkit - an authorized, non-destructive web vulnerability scanner.
- secure-api - a hardened FastAPI backend with a full test suite.
I found and patched a real indirect prompt-injection hole in my own RAG app,
then re-tested until it held.
If you are shipping an AI feature and want it built right, or you already have
one and want someone to try to break it before an attacker does, that is my
lane. I also take on straight full-stack and payment-integration work
(Next.js, FastAPI, Stripe / PromptPay).
I communicate clearly, work async across US, EU, and Asia hours, and I do not
disappear after delivery. Available to start now.
Steps for completing your project
After purchasing the project, send requirements so Phisit can start the project.
Delivery time starts when Phisit receives requirements from you.
Phisit works on your project following the steps below.
Revisions may occur after the delivery date.
Discovery
I confirm the exact scope, your data sources, and how you want the AI to behave.
Setup
I connect your content/data and wire up the AI provider.