Integrations

Connect any agent. See everything.

Native integrations for popular frameworks. SDKs for Python, Node.js, and Go. REST API for everything else. Setup takes minutes.

Join waitlist View quickstart →

Three ways to connect

Choose the integration method that fits your stack

Native Integrations

Pre-built connectors for popular frameworks. Enable in settings, automatic tracking begins.

Recommended for most teams

SDK

Python, Node.js, and Go SDKs with full type support. Maximum control over what gets tracked.

For custom implementations

REST API

Direct API access for any language or platform. Simple POST requests with JSON payloads.

Universal compatibility

Works with your stack

Native support for the tools you already use

AI Providers

OpenAI

GPT-4, Assistants API

Anthropic

Claude models

Google AI

Gemini, PaLM

Cohere

Command, Embed

Mistral

Mistral models

Groq

Fast inference

Agent Frameworks

LangChain

Live in minutes

Install the SDK, add a few lines of code, and every agent action is tracked automatically with full context in xAPI format.

Python SDKpip install empress

Node.js SDKnpm install @empress/sdk

Go SDKgo get github.com/empress/sdk-go

Python

from empress import Empress

empress = Empress(api_key="...")

# Track any agent action
empress.track(
    actor="support-agent",
    verb="resolved",
    object="ticket-892",
    result={"success": True, "time_to_resolve": "4m"},
    context={"customer_tier": "enterprise"}
)

Node.js

import { Empress } from '@empress/sdk';

const empress = new Empress({ apiKey: '...' });

// Track any agent action
await empress.track({
  actor: 'finance-agent',
  verb: 'approved',
  object: 'refund-4892',
  result: { success: true, amount: 127.50 }
});

LangChain Integration

from langchain.callbacks import EmpressCallback
from langchain.chat_models import ChatOpenAI

# Add Empress callback
empress_callback = EmpressCallback(api_key="...")

# Use with any LangChain component
llm = ChatOpenAI(
    model="gpt-4",
    callbacks=[empress_callback]
)

# All LLM calls are automatically tracked
response = llm.invoke("Analyze this customer data...")

# Tracked: actor, verb, object, result, context, cost

Framework Integration

One callback. Complete visibility.

Add the Empress callback to your existing LangChain, LlamaIndex, or CrewAI agents. Every LLM call, tool use, and chain execution is automatically tracked.

Automatic cost tracking per call

Latency and token usage metrics

Full prompt and response capture

Chain execution visualization

Connect your agents today

Native integrations, SDKs, and REST API. Join the beta to get started.

Join waitlist SDK documentation →