@devries-cameron20 / composio Mirror · read-only
Fork 0

Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.

Mirrored from https://github.com/Camo67/composio.git

Synced Nov 21, 06:00 UTC

next
md 77 lines 2.47 KB
Rendered Raw
1 # @composio/langchain
2
3 The LangChain provider turns Composio tools into LangChain `DynamicStructuredTool` objects with built-in execution, ready for LangChain and LangGraph agents. In TypeScript, LangGraph is served by this same package.
4
5 ## Installation
6
7 ```bash
8 npm install @composio/core @composio/langchain @langchain/core @langchain/openai @langchain/langgraph
9 ```
10
11 Set `COMPOSIO_API_KEY` with your API key from [the dashboard](https://dashboard.composio.dev/settings), and `OPENAI_API_KEY` (or your LLM provider's key).
12
13 ## Quickstart
14
15 Create a session for your user, fetch its tools, and wire them into a LangGraph agent:
16
17 ```typescript
18 import { ChatOpenAI } from '@langchain/openai';
19 import { HumanMessage, AIMessage } from '@langchain/core/messages';
20 import { ToolNode } from '@langchain/langgraph/prebuilt';
21 import { StateGraph, MessagesAnnotation } from '@langchain/langgraph';
22 import { Composio } from '@composio/core';
23 import { LangchainProvider } from '@composio/langchain';
24
25 const composio = new Composio({
26 provider: new LangchainProvider(),
27 });
28
29 // Create a session for your user
30 const session = await composio.create('user_123');
31 const tools = await session.tools();
32
33 const toolNode = new ToolNode(tools);
34
35 const model = new ChatOpenAI({
36 model: 'gpt-5.2',
37 temperature: 0,
38 }).bindTools(tools);
39
40 function shouldContinue({ messages }: typeof MessagesAnnotation.State) {
41 const lastMessage = messages[messages.length - 1] as AIMessage;
42 if (lastMessage.tool_calls?.length) {
43 return 'tools';
44 }
45 return '__end__';
46 }
47
48 async function callModel(state: typeof MessagesAnnotation.State) {
49 const response = await model.invoke(state.messages);
50 return { messages: [response] };
51 }
52
53 const workflow = new StateGraph(MessagesAnnotation)
54 .addNode('agent', callModel)
55 .addEdge('__start__', 'agent')
56 .addNode('tools', toolNode)
57 .addEdge('tools', 'agent')
58 .addConditionalEdges('agent', shouldContinue);
59
60 const app = workflow.compile();
61
62 const finalState = await app.invoke({
63 messages: [
64 new HumanMessage(
65 "Send an email to john@example.com with the subject 'Hello' and body 'Hello from Composio!'"
66 ),
67 ],
68 });
69 console.log(finalState.messages[finalState.messages.length - 1].content);
70 ```
71
72 Each tool is a standard `DynamicStructuredTool`, so it also works anywhere LangChain accepts tools: chains, LCEL pipelines, and `bindTools` on any chat model.
73
74 ## Links
75
76 - [LangChain provider docs](https://docs.composio.dev/docs/providers/langchain)
77 - [Composio documentation](https://docs.composio.dev)