Integrations / AutoGen
Launch and agentsDocumented
Connect AutoGen workflows to Badgr
Use Badgr as the model for AutoGen agents, and wrap Badgr jobs as a tool for heavy steps.
The steps follow the tool's documented settings. No tested run is recorded for this page yet.
Why use Badgr for AutoGen
- Badgr interface
- AI API
- Where AutoGen fits
- External orchestrator
- Why Badgr
- Models and execution tools
- The OpenAI-compatible client works against Badgr, so agents keep their code.
- Heavy steps can run as a capped job through a tool.
- One account covers model calls and compute.
What you can do
- A multi-agent conversation backed by Badgr models
- An agent that starts a GPU job and reads the result
- Compare Badgr models inside one team of agents
Quickstart: A model client for Badgr
import os
from autogen_ext.models.openai import OpenAIChatCompletionClient
model_client = OpenAIChatCompletionClient(
model="badgr-auto",
base_url="https://aibadgr.com/v1",
api_key=os.environ["BADGR_API_KEY"],
model_info={
"vision": False,
"function_calling": True,
"json_output": True,
"family": "unknown",
},
)How it works
- 1.Your agents use the Badgr model client
- 2.Replies from Badgr drive the conversation
- 3.For heavy steps, a tool starts a capped job
- 4.The tool returns the result to the agents
What you get
- Agent conversations answered by Badgr models
- Job results, if you add the job tool
Requirements
- A Badgr API key, saved as BADGR_API_KEY (get one at /get-api-key)
- Python and the autogen-ext package with its OpenAI extra
- For jobs: the Badgr CLI and a funded balance
What it costs
- Model calls are billed per request at the model's rate (see Models).
- Jobs are billed for run time and bounded by --max-cost. See the LangGraph or CrewAI page for a tool example.
See pricing for how billing works.
Proof
No tested run is recorded for this integration yet. A receipt and example output will be added here after a real run, not before.
Troubleshooting
Unknown model error
AutoGen needs model_info for names it does not know. Keep the block shown above.
Function calling fails
Not every model supports tools. Set function_calling to False if yours does not.
Agents never stop
Set a maximum number of turns or a termination condition.
Developer reference
Ready to start with AutoGen?
Every run or service has a spending limit you set before it starts.