I will build an ai chatbot for whatsapp or your website


About this gig
I build AI agents that run in production, not demo notebooks.
I have shipped agents on the Anthropic Claude API: a security-training agent, a marketing content agent, a Notion organizer, and internal assistants that query real databases. I work with Claude, GPT and Gemini.
WHAT YOU GET
- A working AI agent in Node.js/TypeScript, deployed and running
- Tool use / function calling wired to your real systems: REST APIs, PostgreSQL, Notion, WhatsApp, Discord
- Prompt architecture, error handling, retries and token-cost control
- Clean source code, README and a handover call
TYPICAL USE CASES
- Support agent that reads your knowledge base
- Internal agent that queries your database in plain language
- Invoice and document extraction
- Scheduled content or report generation
HOW I WORK
Before writing code I confirm scope in writing: inputs, outputs, integrations, and what done means. No scope drift.
17 years in IT and I deploy everything myself (Linux, nginx, PM2, Docker, CI/CD), so your agent keeps running after delivery.
Message me before ordering so I can confirm scope and quote accurately.
Get to know Hector Mejia
Software Product Lead
- FromMexico
- Member sinceAug 2026
- Avg. response time1 hour
Languages
English, Spanish
FAQ
Do I need my own Claude or OpenAI API key?
Yes, and that is intentional. The account stays in your name so you own the agent and control the billing. I help you set it up during the project and you can revoke my access whenever you want.
Can the agent connect to my database or internal API?
Yes, that is the core of what I do. I connect agents to PostgreSQL, MySQL, REST APIs, Notion, WhatsApp and Discord using tool calling. Send me your docs before ordering and I will confirm feasibility.
Where does the agent run after you deliver it?
On your own server or cloud account, never mine. I handle the full deployment: Linux, nginx, PM2, SSL and process monitoring. You get the credentials, the source code and documentation to manage it yourself.
How much will the API cost me per month?
It depends on your volume, so I estimate it before we start based on your expected usage. I also build in token-cost controls, caching and model selection so the bill stays predictable instead of surprising you.
Can you fine-tune a model on my own data?
I do not offer fine-tuning. For almost every business case, retrieval over your own documents gives better results at a fraction of the cost, and that I do include. If you truly need fine-tuning, I will tell you honestly.

