Prerequisites
- A running Langflow instance (local, Docker, or self-hosted)
- A Qhaigc API key — get one from the API Tokens page
- For RAG flows: an embedding component configured separately
Configuration Steps
Add the OpenAI text generation component
From the component library, drag in the OpenAI text generation component from the OpenAI bundle. This is your primary chat/LLM node.
Fill in the text component parameters
Configure the following fields in the component panel:
| Field | Value |
|---|---|
| OpenAI API Key | Your Qhaigc API key (starts with sk-) |
| Model Name | e.g. gpt-4o, gpt-4o-mini, or another model |
| Temperature | Your preferred value (e.g. 0.7) |
| Max Tokens | Set as needed |
Langflow’s OpenAI text generation component may not expose a Base URL field in the current version. Configure the fields listed above; refer to Langflow’s OpenAI bundle documentation for the exact fields available in your version.
Add an OpenAI Embeddings component for RAG flows
If your flow includes a knowledge base or vector store, drag in a separate OpenAI Embeddings component and configure it:
The embedding component does expose an API base field that accepts a custom URL.
| Field | Value |
|---|---|
| OpenAI API Key | Your Qhaigc API key |
| OpenAI API Base | https://api.qhaigc.net/v1 |
| Model | e.g. bge-m3 or text-embedding-3-large |
Connect input and output components
Wire the Prompt, Memory, Vector Store, or Chat Output nodes to your main model node as needed by your flow.
Verifying the Connection
Your setup is working when:- The text generation component runs and returns a reply from Qhaigc
- If you configured embeddings, the vectorization and retrieval steps complete without errors
Frequently Asked Questions
Why does the text component not have a Base URL field? Langflow’s OpenAI bundle text generation component listsapi_key, model, max_tokens, and temperature as its primary fields. A custom Base URL field may or may not be present depending on your Langflow version. Check your version’s component panel or the official docs for confirmation.
Why do I need a separate embedding component for RAG?
Langflow’s knowledge base and retrieval chain requires an independent embedding component — the text generation component does not provide embeddings automatically.