agenttrafficlab

Minimal zero-dependency Python client for Agent Traffic Lab.

Install

pip install agenttrafficlab

Recommended usage: choose + execute + automatic Outcome

from agenttrafficlab import run

result = run("summarize a support ticket for handoff")
print(result["execution"]["output_text"])
print(result["execution"]["outcome_status"])

run() asks ATL to choose the provider, executes the Decision-authorized provider through ATL, uses bounded fallback when needed, and records the durable execution Outcome automatically. The caller does not need to submit a separate feedback call for ATL-owned execution.

Decision only

from agenttrafficlab import decide

result = decide("summarize a support ticket for handoff")
print(result["recommended_provider"]["provider_id"])

Use decide() when your own system will execute the provider outside ATL. In that mode ATL cannot observe the external execution automatically, so the execution result still needs to be reported if you want it included as Outcome evidence.

Provider selection remains ATL-owned; callers do not choose a provider directly.

Public surfaces:

• Decision API: https://mcp.agenttrafficlab.com/intent-entrance/decision
• Execution API: https://mcp.agenttrafficlab.com/intent-entrance/execute
• MCP endpoint: https://mcp.agenttrafficlab.com/mcp
• Live demo: https://mcp.agenttrafficlab.com/try

Python 3.9+; no runtime dependencies.