I will integrate chatgpt or claude API into your existing app


About this gig
Do you want ChatGPT or Claude inside your own app? Not a demo. A working feature your users actually touch.
I connect the API to software you already have. Web app, desktop tool, internal system, or a backend script.
WHAT I DO:
- Wire the API into your existing codebase, no rewrite
- Design and test the prompts so the output stays consistent
- Handle what breaks in production: rate limits, timeouts, malformed responses, token cost
- Add streaming, conversation memory, or document handling if you need it
WHY ME:
I built and shipped a production document-processing system currently used in pharmacies. Windows desktop app, mobile apps, backend, and the AI pipeline. All built solo. I know what a model does when a real user feeds it a real mess.
These are working integrations, not demos.
I work in writing and give clear updates. No surprises, no disappearing.
Send me your stack and what you want the AI to do. I will tell you honestly whether the API is the right answer for your case.
Get to know Harun Yorulmaz
Python Developer , AI and LLM Integration , Desktop and Mobile Apps
- FromTurkey
- Member sinceAug 2026
Languages
Turkish, English
FAQ
Do I need to give you my whole codebase?
No. I need the part that will call the API and enough context to run it. Many jobs are done with a single module and a test environment. If you cannot share code at all, I deliver a standalone service your developer plugs in.
Who pays for the API usage?
You do, on your own OpenAI or Anthropic account. I never leave my key in your product. I estimate your monthly cost before we start so there are no surprises later.
What if my app is not written in Python?
Fine. The API is HTTP, so it works from any language. I write the integration layer in Python or as a service your app calls. Tell me your stack in a message and I will confirm the approach before you order.
Can you handle confidential data?
Yes. I built a system that processes pharmacy documents under strict privacy rules. I can design it so nothing sensitive leaves your own servers. Tell me this upfront, it changes the architecture.
How accurate will the answers be?
No language model is 100 percent correct, and anyone promising that is selling you something. I build validation around the model: checked output formats, fallbacks when it fails, and low confidence answers flagged instead of hidden.
What happens when the model or the API changes?
I keep the model behind a small layer, so swapping ChatGPT for Claude or moving to a newer version is a config change, not a rewrite. Within the revision window I fix anything that breaks.
Can you work with my existing developer?
Yes, and it usually makes the job faster. I work in writing, leave readable code, and hand over a short setup note with every delivery.

