AutoGen (gateway)

Route AutoGen model calls through the Respan gateway to use 250+ models. Only your Respan API key is needed — no separate provider key is required in the example process.

Setup

Install packages

$ pip install respan-ai respan-instrumentation-autogen

Set environment variables

$ export RESPAN_API_KEY="YOUR_RESPAN_API_KEY"
$ export RESPAN_BASE_URL="https://api.respan.ai/api"
$ export RESPAN_MODEL="gpt-4o-mini"

Point AutoGen to the Respan gateway

1 import os
2 
3 from autogen_ext.models.openai import OpenAIChatCompletionClient
4 
5 model_client = OpenAIChatCompletionClient(
6     model=os.getenv("RESPAN_MODEL", "gpt-4o-mini"),
7     api_key=os.environ["RESPAN_API_KEY"],
8     base_url=os.getenv("RESPAN_BASE_URL", "https://api.respan.ai/api"),
9     model_info={
10         "vision": False,
11         "function_calling": True,
12         "json_output": True,
13         "structured_output": True,
14         "family": "unknown",
15     },
16 )

Switch models

Change RESPAN_MODEL to use a different model through the same gateway.

$ export RESPAN_MODEL="claude-sonnet-4-5-20250929"

See the full model list.

$	export RESPAN_API_KEY="YOUR_RESPAN_API_KEY"
$	export RESPAN_BASE_URL="https://api.respan.ai/api"
$	export RESPAN_MODEL="gpt-4o-mini"

1	import os
2
3	from autogen_ext.models.openai import OpenAIChatCompletionClient
4
5	model_client = OpenAIChatCompletionClient(
6	model=os.getenv("RESPAN_MODEL", "gpt-4o-mini"),
7	api_key=os.environ["RESPAN_API_KEY"],
8	base_url=os.getenv("RESPAN_BASE_URL", "https://api.respan.ai/api"),
9	model_info={
10	"vision": False,
11	"function_calling": True,
12	"json_output": True,
13	"structured_output": True,
14	"family": "unknown",
15	},
16	)