Skip to main content

Choose Your Integration

cascadeflow integrates with every major agent framework. Pick the one that matches your stack.

Quick Decision

If you use…InstallIntegration typeGuide
LangChain / LangGraphpip install cascadeflow[langchain]Callback handlerLangChain →
OpenAI Agents SDKpip install cascadeflow[openai-agents]ModelProviderOpenAI Agents →
CrewAIpip install cascadeflow[crewai]llm_hooksCrewAI →
Google ADKpip install cascadeflow[google-adk]BasePluginGoogle ADK →
Vercel AI SDKnpm install @cascadeflow/vercel-aiMiddlewareVercel AI →
n8nnpm install @cascadeflow/n8n-nodes-cascadeflowCommunity noden8n →
Direct SDK callspip install cascadeflowinit() + run()Observe →
TypeScript (no framework)npm install @cascadeflow/coreCascadeAgentCore →

By Language

Python

All Python integrations share the same Harness API — init(), run(), @agent(). The framework integration handles the bridge:
# LangChain — callback handler
from cascadeflow.integrations.langchain import CascadeFlowHandler
handler = CascadeFlowHandler()
chain.invoke(query, config={"callbacks": [handler]})

# OpenAI Agents — model provider
from cascadeflow.integrations.openai_agents import CascadeFlowModelProvider
provider = CascadeFlowModelProvider()

# CrewAI — hooks
from cascadeflow.integrations.crewai import CascadeFlowHooks
hooks = CascadeFlowHooks()

# Google ADK — plugin
from cascadeflow.integrations.google_adk import CascadeFlowPlugin
plugin = CascadeFlowPlugin()

TypeScript

// Core (standalone)
import { CascadeAgent } from '@cascadeflow/core';

// Vercel AI SDK (middleware)
import { createChatHandler } from '@cascadeflow/vercel-ai';

// LangChain (withCascade)
import { withCascade } from '@cascadeflow/langchain';

Which One If You Are Not Sure?

All Integrations

For a full comparison matrix including feature coverage per framework, see Integrations Overview →

You’re Ready

You’ve completed the Getting Started path. From here: