Browse categories
Explore
Fiverr Pro
English
$
USD
I will test your mcp server and ai agent tool calls with automated evals
QA Automation Automation Tester Playwright Framework
I help startups and businesses release reliable AI products.
I specialize in Agentic AI Testing, LLM Evaluation, RAG Testing, and AI QA for chatbots, AI assistants, and tool-using agents.
What I tes...
About this Gig
Launching an AI agent or MCP server? Test it before launch!
AI agents call the wrong tool, pass bad arguments, or fail silently. I catch these problems early with structured AI agent testing, MCP server testing, and automated LLM evaluation.
ABOUT ME
I'm a Agentic AI/LLM automation engineer with 4+ years of experience, including validating AI model outputs in production analytics. I turn vague bugs into clear, repeatable tests.
WHAT I TEST
- Tool calling accuracy and argument validation
- MCP server schemas, error handling, and edge cases
- Task completion, infinite loops, and failed steps
- Prompt injection and unsafe action checks
- Response consistency across repeated runs
WHAT YOU GET
- Test scenarios built around your use case
- ️ Automated eval scripts you can rerun anytime
- GitHub Actions CI that flags regressions (Premium)
- Prioritized bug report with reproduction steps
WORKS WITH
MCP servers, LangChain, LangGraph, CrewAI, n8n, and custom agents.
Tools: Python, TypeScript, Promptfoo, DeepEval.
Private code, NDA available. Sandbox keys only.
Message me before ordering with your agent type, and I'll confirm it's a fit!
Testing application:
API
Development technology:
JavaScript
•
Node.js
•
Python
•
TypeScript
•
Other
Device:
PC
•
Mac
•
Linux
FAQ
Question: Do you build the AI agent or MCP server?
Answer: No, I test and evaluate it. My report includes recommended fixes, so your developer can resolve issues quickly.
Question: Which frameworks and tools do you support?
Answer: MCP servers, LangChain, LangGraph, CrewAI, n8n and custom agents or APIs. I write tests in Python and TypeScript using Promptfoo and DeepEval.
Question: Do you need my API keys or production access?
Answer: Only sandbox or test keys. I never need production secrets, so please do not share them.
Question: Will you fix the bugs you find?
Answer: The report lists each bug with reproduction steps and a recommended fix. Implementing the fixes is a separate order.
Question: Is my code kept private?
Answer: Yes. Your code and data stay confidential, and I’m happy to sign an NDA before starting.
Question: What do I need to send before ordering?
Answer: Message me with your agent type, framework and model, plus a repo, endpoint or demo link. I’ll confirm it’s a good fit within a few hours.
