Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.
docs: add LangGraph and Cloudflare Workers AI provider pages
LangGraph: Python-only, uses create_agent + invoke pattern. Cloudflare: TS-only (JS code block since @cloudflare/workers-types crashes the docs type checker), Workers AI function calling with agentic loop. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Sushmithamallesh committed
Feb 17, 2026 at 17:52 UTC
fba1279ffa9c548de10584eb92b4c0fbd6ce174d
3 files changed
+154
docs/content/docs/providers/cloudflare.mdx
new
+99
@@ -0,0 +1,99 @@
1
+---
2
+title: Cloudflare Workers AI
3
+description: Build an agent with Cloudflare Workers AI and enable it to use 800+ tools with Composio.
4
+keywords: [cloudflare, workers ai]
5
+---
6
+
7
+The Cloudflare provider transforms Composio tools into Cloudflare Workers AI's [function calling format](https://developers.cloudflare.com/workers-ai/function-calling/) with manual tool execution.
8
+
9
+<Steps>
10
+<Step>
11
+**Install**
12
+
13
+```bash
14
+npm install @composio/core @composio/cloudflare
15
+```
16
+</Step>
17
+
18
+<Step>
19
+**Configure API Keys**
20
+
21
+<Callout type="info">
22
+Set `COMPOSIO_API_KEY` with your API key from [Settings](https://platform.composio.dev/?next_page=/settings). Your Cloudflare Worker has access to AI via the `env.AI` binding.
23
+</Callout>
24
+
25
+Add `COMPOSIO_API_KEY` to your Worker's environment variables in `wrangler.toml` or via the Cloudflare dashboard.
26
+
27
+```toml title="wrangler.toml"
28
+[vars]
29
+COMPOSIO_API_KEY = "xxxxxxxxx"
30
+```
31
+</Step>
32
+
33
+<Step>
34
+**Create session and run**
35
+
36
+```js
37
+import { Composio } from "@composio/core";
38
+import { CloudflareProvider } from "@composio/cloudflare";
39
+
40
+export default {
41
+ async fetch(request, env) {
42
+ const composio = new Composio({
43
+ apiKey: env.COMPOSIO_API_KEY,
44
+ provider: new CloudflareProvider(),
45
+ });
46
+
47
+ // Create a session for your user
48
+ const session = await composio.create("user_123");
49
+ const tools = await session.tools();
50
+
51
+ const toolsArray = Object.values(tools);
52
+
53
+ let response = await env.AI.run("@cf/meta/llama-4-scout-17b-16e-instruct", {
54
+ messages: [
55
+ {
56
+ role: "user",
57
+ content:
58
+ "Send an email to john@example.com with the subject 'Hello' and body 'Hello from Composio!'",
59
+ },
60
+ ],
61
+ tools: toolsArray,
62
+ });
63
+
64
+ // Agentic loop — keep executing tool calls until the model responds with text
65
+ while (response.tool_calls && response.tool_calls.length > 0) {
66
+ const toolResults = [];
67
+ for (const toolCall of response.tool_calls) {
68
+ const result = await composio.provider.executeToolCall(
69
+ "user_123",
70
+ { name: toolCall.name, arguments: toolCall.arguments },
71
+ {}
72
+ );
73
+ toolResults.push({
74
+ role: "tool",
75
+ tool_call_id: toolCall.name,
76
+ content: result,
77
+ });
78
+ }
79
+
80
+ response = await env.AI.run("@cf/meta/llama-4-scout-17b-16e-instruct", {
81
+ messages: [
82
+ {
83
+ role: "user",
84
+ content:
85
+ "Send an email to john@example.com with the subject 'Hello' and body 'Hello from Composio!'",
86
+ },
87
+ { role: "assistant", content: "", tool_calls: response.tool_calls },
88
+ ...toolResults,
89
+ ],
90
+ tools: toolsArray,
91
+ });
92
+ }
93
+
94
+ return new Response(response.response ?? "No response");
95
+ },
96
+};
97
+```
98
+</Step>
99
+</Steps>
docs/content/docs/providers/langgraph.mdx
new
+53
@@ -0,0 +1,53 @@
1
+---
2
+title: LangGraph
3
+description: Build an agent with LangGraph and enable it to use 800+ tools with Composio.
4
+keywords: [langgraph]
5
+---
6
+
7
+The LangGraph provider transforms Composio tools into LangChain's [StructuredTool](https://python.langchain.com/docs/how_to/custom_tools/) format for use with [LangGraph](https://langchain-ai.github.io/langgraph/) agents.
8
+
9
+<Steps>
10
+<Step>
11
+**Install**
12
+
13
+```bash
14
+pip install composio composio_langgraph langgraph langchain_openai
15
+```
16
+</Step>
17
+
18
+<Step>
19
+**Configure API Keys**
20
+
21
+<Callout type="info">
22
+Set `COMPOSIO_API_KEY` with your API key from [Settings](https://platform.composio.dev/?next_page=/settings) and `OPENAI_API_KEY` with your [OpenAI API key](https://platform.openai.com/api-keys).
23
+</Callout>
24
+
25
+```txt title=".env"
26
+COMPOSIO_API_KEY=xxxxxxxxx
27
+OPENAI_API_KEY=xxxxxxxxx
28
+```
29
+</Step>
30
+
31
+<Step>
32
+**Create session and run**
33
+
34
+```python
35
+from composio import Composio
36
+from composio_langgraph import LanggraphProvider
37
+from langchain.agents import create_agent
38
+from langchain_openai import ChatOpenAI
39
+
40
+composio = Composio(provider=LanggraphProvider())
41
+llm = ChatOpenAI(model="gpt-5.2")
42
+
43
+# Create a session for your user
44
+session = composio.create(user_id="user_123")
45
+tools = session.tools()
46
+
47
+agent = create_agent(tools=tools, model=llm)
48
+result = agent.invoke({"messages": [("user", "Send an email to john@example.com with the subject 'Hello' and body 'Hello from Composio!'")]})
49
+
50
+print(result["messages"][-1].content)
51
+```
52
+</Step>
53
+</Steps>
docs/content/docs/providers/meta.json
+2
@@ -5,11 +5,13 @@
5
"anthropic",
6
"vercel",
7
"langchain",
8
+ "langgraph",
9
"mastra",
10
"openai",
11
"google",
12
"llamaindex",
13
"crewai",
14
+ "cloudflare",
15
"custom-providers"
16
]
17
}