Use the OpenAI Responses API with Respan for logging and observability

Overview

The Respan gateway supports OpenAI's Responses API, enabling automatic logging and observability for text inputs, image inputs, file search, function calling, web search, and streaming responses.

The integration works as a pass-through - modify your client's base URL and pass Respan parameters via a custom header. All standard Responses API features are supported including reasoning models and stateful interactions.

How it works

Point your OpenAI client to the Respan gateway base URL (https://api.keywordsai.co/api) and use your Respan API key for authentication.

Pass Respan-specific parameters like customer_identifier and metadata through the X-Data-Keywordsai-Params header as base64-encoded JSON. Every request is logged with full observability.

python

from openai import OpenAI
import json, base64

client = OpenAI(
    base_url="https://api.keywordsai.co/api",
    api_key="YOUR_RESPAN_API_KEY",
)

# Pass Respan params via header
params = json.dumps({"customer_identifier": "user-123"})
encoded = base64.b64encode(params.encode()).decode()

response = client.responses.create(
    model="gpt-4o",
    input="Explain quantum computing",
    extra_headers={"X-Data-Keywordsai-Params": encoded},
)