CrewAI (gateway)

Route CrewAI’s underlying LLM calls through the Respan gateway to use 250+ models from different providers. Only your RESPAN_API_KEY is needed - no separate provider keys required.

Setup

Install packages

$ pip install crewai

Set environment variables

$ export RESPAN_API_KEY="YOUR_RESPAN_API_KEY"

No separate OpenAI provider key is needed. The code below points CrewAI’s OpenAI-compatible traffic to the Respan gateway.

Point CrewAI to the Respan gateway

1 import os
2 
3 from crewai import Agent, Crew, LLM, Task
4 
5 respan_api_key = os.environ["RESPAN_API_KEY"]
6 respan_base_url = os.getenv("RESPAN_BASE_URL", "https://api.respan.ai/api")
7 
8 os.environ["OPENAI_API_KEY"] = respan_api_key
9 os.environ["OPENAI_BASE_URL"] = respan_base_url
10 
11 llm = LLM(
12     model="gpt-5.5",
13     api_key=respan_api_key,
14     base_url=respan_base_url,
15 )
16 
17 researcher = Agent(
18     role="Researcher",
19     goal="Research the latest AI trends",
20     backstory="Senior AI researcher.",
21     llm=llm,
22 )
23 
24 task = Task(
25     description="Research the latest trends in AI agent frameworks.",
26     expected_output="A summary of key trends.",
27     agent=researcher,
28 )
29 
30 crew = Crew(agents=[researcher], tasks=[task])
31 print(crew.kickoff())

Switch models

Change the model parameter on LLM to use 250+ models from different providers through the same gateway.

1 llm = LLM(
2     model="claude-sonnet-4-5-20250929",
3     api_key=respan_api_key,
4     base_url=respan_base_url,
5 )

See the full model list.