layerinfinite-sdk 0.3.0

Decision intelligence layer for AI agents — Layerinfinite

Layerinfinite Python SDK

Layerinfinite is a decision intelligence middleware for AI agents — it records action outcomes, computes composite trust scores, and recommends the highest-performing next action so your agents learn from every decision. Drop it between any LLM agent and your production infrastructure in minutes.

Installation

pip install layerinfinite-sdk

Quick Start

from layerinfinite import LayerinfiniteClient, LogOutcomeRequest

client = LayerinfiniteClient(api_key="layerinfinite_your_key")

# Ask Layerinfinite which action to take
scores = client.get_scores(
    agent_id="my-agent",
    issue_type="billing_dispute"
)
print(scores.top_action.action_name)   # e.g. "escalate_to_senior"
print(f"Policy: {scores.policy}")      # exploit | explore | escalate

# Log the outcome after the action runs
client.log_outcome(LogOutcomeRequest(
    agent_id="my-agent",
    action_id=scores.top_action.action_id,
    context_id=scores.context_id,
    issue_type="billing_dispute",
    success=True,
    outcome_score=0.9,
    business_outcome="resolved"
))

Context Manager

with LayerinfiniteClient(api_key="layerinfinite_your_key") as client:
    scores = client.get_scores(agent_id="agent-1", issue_type="payment_failed")
    # Session is automatically closed on exit

LangChain Integration

from layerinfinite import LayerinfiniteClient, LogOutcomeRequest
from langchain_core.tools import tool

layerinfinite = LayerinfiniteClient(api_key="layerinfinite_your_key")

@tool
def resolve_ticket(agent_id: str, issue_type: str) -> str:
    """Resolve a support ticket using Layerinfinite-guided action."""
    scores = layerinfinite.get_scores(agent_id=agent_id, issue_type=issue_type)
    action = scores.top_action.action_name
    # ... run the action ...
    layerinfinite.log_outcome(LogOutcomeRequest(
        agent_id=agent_id,
        action_id=scores.top_action.action_id,
        context_id=scores.context_id,
        issue_type=issue_type,
        success=True, outcome_score=0.85, business_outcome="resolved"
    ))
    return action

Error Handling

from layerinfinite import LayerinfiniteClient, LayerinfiniteAuthError, LayerinfiniteRateLimitError
import time

client = LayerinfiniteClient(api_key="layerinfinite_your_key")
try:
    scores = client.get_scores(agent_id="agent-1", issue_type="billing")
except LayerinfiniteAuthError:
    print("Invalid API key — check LAYERINFINITE_API_KEY")
except LayerinfiniteRateLimitError as e:
    print(f"Rate limited — retry after {e.retry_after}s")
    time.sleep(e.retry_after)

Configuration

Parameter	Default	Description
api_key	required	Your Layerinfinite API key
base_url	https://your-app.railway.app	API base URL
timeout	10.0	Request timeout in seconds
max_retries	3	Max retries on 429/5xx

Links

• PyPI: pypi.org/project/layerinfinite-sdk
• Docs: docs.layerinfinite.ai
• GitHub: github.com/hari08varma/Outcome