Skip to main content

OpenAI Collective

A collective for developers utilizing OpenAI's foundational models and APIs to build, integrate, and innovate. The OpenAI Collective is organized and managed by the Stack Overflow community as a resource for developers.
Contact

Can you answer these questions?

View all unanswered questions

These questions still don't have an answer

0 votes
0 answers
60 views

VS Code Codex extension fails with "spawn UNKNOWN" on Windows even though `codex app-server` works

I’m trying to use the OpenAI Codex extension in VS Code on Windows, but the extension fails immediately with: Failed to start Codex process: spawn UNKNOWN Context: - Windows machine - enterprise-...
2 votes
0 answers
184 views
+50

OpenAI RAG LangChain tool calling is not working/compatible with withStructuredOutput()?

I have a OpenAI model with Retrieval-Augmented Generation (RAG): import {OpenAIEmbeddingFunction} from "@chroma-core/openai"; import chromaClient from "../config/chromadb"; import {...
Tooling
0 votes
0 replies
45 views

Looking for a reliable and cost-effective token relay for GPT/Claude/Gemini API

I'm developing an AI application that uses GPT-5.4 and Claude-sonnet-4-6. The official API costs are too high for my scale (thousands of requests per day). I've heard about token relay services that ...
-2 votes
0 answers
59 views

How to store streaming AI responses per message in React without them merging together?

I'm currently trying to write a my first React project using AI API and it's okay for now, except for one thing: every time I send a message, I use use state to display it, but by doing that, if I ...
Best practices
1 vote
0 replies
62 views

Need to design a pdf batch file processsing system

I get around 100,000 pdf files (1-50MB each) a month. We get file IDs, pull and parse files, build chunks, embed them, and store documents + vectors for semantic search. What can be the best design ...

Looking for an extra challenge?

View all bountied questions

These questions have a bounty on them

2 votes
0 answers
184 views
+50

OpenAI RAG LangChain tool calling is not working/compatible with withStructuredOutput()?

I have a OpenAI model with Retrieval-Augmented Generation (RAG): import {OpenAIEmbeddingFunction} from "@chroma-core/openai"; import chromaClient from "../config/chromadb"; import {...