volcano-agent-sdk

Name: volcano-agent-sdk
Author: Kong

108installs/mo · npm399stars32forks

How to install volcano-agent-sdk?

kong/volcano-agent-sdk

$git clone https://github.com/Kong/volcano-agent-sdk

Installs into the current project.

Summary

The TypeScript SDK for Multi-Provider AI Agents

Files · 1

View on GitHub

README.md

1	[![CI](https://github.com/Kong/volcano-agent-sdk/actions/workflows/ci.yml/badge.svg)](https://github.com/Kong/volcano-agent-sdk/actions/workflows/ci.yml)
2	[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)
3	[![npm](https://img.shields.io/npm/v/@volcano.dev/agent.svg)](https://www.npmjs.com/package/@volcano.dev/agent)
4
5	# 🌋 Volcano Agent SDK
6
7	The TypeScript SDK for Multi-Provider AI Agents
8
9	Build agents that chain LLM reasoning with MCP tools. Mix OpenAI, Claude, Mistral in one workflow. Parallel execution, branching, loops. Native retries, streaming, and typed errors.
10
11	📚 [Read the full documentation at volcano.dev →](https://volcano.dev/)
12
13	## ✨ Features
14
15	<table>
16	<tr>
17	<td width="33%">
18
19	### 🤖 Automatic Tool Selection
20	LLM automatically picks which MCP tools to call based on your prompt. No manual routing needed.
21
22	</td>
23	<td width="33%">
24
25	### 🧩 Multi-Agent Crews
26	Define specialized agents and let the coordinator autonomously delegate tasks. Like automatic tool selection, but for agents.
27
28	</td>
29	<td width="33%">
30
31	### 💬 Conversational Results
32	Ask questions about what your agent did. Use `.summary()` or `.ask()` instead of parsing JSON.
33
34	</td>
35	</tr>
36
37	<tr>
38	<td width="33%">
39
40	### 🔧 100s of Models
41	OpenAI, Anthropic, Mistral, Bedrock, Vertex, Azure. Switch providers per-step or globally.
42
43	</td>
44	<td width="33%">
45
46	### 🔄 Advanced Patterns
47	Parallel execution, branching, loops, sub-agent composition. Enterprise-grade workflow control.
48
49	</td>
50	<td width="33%">
51
52	### 📡 Streaming
53	Stream tokens in real-time as LLMs generate them. Perfect for chat UIs and SSE endpoints.
54
55	</td>
56	</tr>
57
58	<tr>
59	<td width="33%">
60
61	### 🛡️ TypeScript-First
62	Full type safety with IntelliSense. Catch errors before runtime.
63
64	</td>
65	<td width="33%">
66
67	### 📊 Observability
68	OpenTelemetry traces and metrics. Export to Jaeger, Prometheus, DataDog, or any OTLP backend.
69
70	</td>
71	<td width="33%">
72
73	### ⚡ Production-Ready
74	Built-in retries, timeouts, error handling, and connection pooling. Battle-tested at scale.
75
76	</td>
77	</tr>
78	</table>
79
80	[Explore all features →](https://volcano.dev/docs#key-features)
81
82	## Quick Start
83
84	### Installation
85
86	```bash
87	npm install @volcano.dev/agent
88	```
89
90	That's it! Includes MCP support and all common LLM providers (OpenAI, Anthropic, Mistral, Llama, Vertex).
91
92	[View installation guide →](https://volcano.dev/docs#installation)
93
94	### Hello World with Automatic Tool Selection
95
96	```ts
97	import { agent, llmOpenAI, mcp } from "@volcano.dev/agent";
98
99	const llm = llmOpenAI({
100	apiKey: process.env.OPENAI_API_KEY!,
101	model: "gpt-4o-mini"
102	});
103
104	const weather = mcp("http://localhost:8001/mcp");
105	const tasks = mcp("http://localhost:8002/mcp");
106
107	// Agent automatically picks the right tools
108	const results = await agent({ llm })
109	.then({
110	prompt: "What's the weather in Seattle? If it will rain, create a task to bring an umbrella",
111	mcps: [weather, tasks] // LLM chooses which tools to call
112	})
113	.run();
114
115	// Ask questions about what happened
116	const summary = await results.summary(llm);
117	console.log(summary);
118	```
119
120	### Multi-Agent Coordinator
121
122	```ts
123	import { agent, llmOpenAI } from "@volcano.dev/agent";
124
125	const llm = llmOpenAI({ apiKey: process.env.OPENAI_API_KEY! });
126
127	// Define specialized agents
128	const researcher = agent({ llm, name: 'researcher', description: 'Finds facts and data' })
129	.then({ prompt: "Research the topic." })
130	.then({ prompt: "Summarize the research." });
131
132	const writer = agent({ llm, name: 'writer', description: 'Creates content' })
133	.then({ prompt: "Write content." });
134
135	// Coordinator autonomously delegates to specialists
136	const results = await agent({ llm })
137	.then({
138	prompt: "Write a blog post about quantum computing",
139	agents: [researcher, writer] // Coordinator decides when done
140	})
141	.run();
142
143	// Ask what happened
144	const post = await results.ask(llm, "Show me the final blog post");
145	console.log(post);
146	```
147
148	[View more examples →](https://volcano.dev/docs/examples)
149
150	## Documentation
151
152	### 📖 Comprehensive Guides

People who liked this also viewed

6 picks

Type

fastmcp🚀 The fast, Pythonic way to build MCP servers and clients.MCP ServersJul 202686M27k
playwright-mcpPlaywright MCP serverMCP ServersJul 202626M35k
ext-appsOfficial repo for spec & SDK of MCP Apps protocol - standard for UIs embedded AI chatbots, served by MCP serversMCP ServersJul 20268.5M2.6k
langchain-mcp-adaptersLangChain 🔌 MCPMCP ServersJul 20267.0M3.6k
mcporterCall MCPs via TypeScript, masquerading as simple TypeScript API. Or package them as cli.MCP ServersJul 20262.5M4.8k
mcp-proxyA bridge between Streamable HTTP and stdio MCP transportsMCP ServersJul 20262.2M2.7k

volcano-agent-sdk

108installs/mo · npm399stars32forks

View on GitHub

How to install volcano-agent-sdk?

kong/volcano-agent-sdk

$git clone https://github.com/Kong/volcano-agent-sdk

Installs into the current project.

Summary

The TypeScript SDK for Multi-Provider AI Agents

Files · 1

View on GitHub

README.md

1	[![CI](https://github.com/Kong/volcano-agent-sdk/actions/workflows/ci.yml/badge.svg)](https://github.com/Kong/volcano-agent-sdk/actions/workflows/ci.yml)
2	[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE)
3	[![npm](https://img.shields.io/npm/v/@volcano.dev/agent.svg)](https://www.npmjs.com/package/@volcano.dev/agent)
4
5	# 🌋 Volcano Agent SDK
6
7	The TypeScript SDK for Multi-Provider AI Agents
8
9	Build agents that chain LLM reasoning with MCP tools. Mix OpenAI, Claude, Mistral in one workflow. Parallel execution, branching, loops. Native retries, streaming, and typed errors.
10
11	📚 [Read the full documentation at volcano.dev →](https://volcano.dev/)
12
13	## ✨ Features
14
15	<table>
16	<tr>
17	<td width="33%">
18
19	### 🤖 Automatic Tool Selection
20	LLM automatically picks which MCP tools to call based on your prompt. No manual routing needed.
21
22	</td>
23	<td width="33%">
24
25	### 🧩 Multi-Agent Crews
26	Define specialized agents and let the coordinator autonomously delegate tasks. Like automatic tool selection, but for agents.
27
28	</td>
29	<td width="33%">
30
31	### 💬 Conversational Results
32	Ask questions about what your agent did. Use `.summary()` or `.ask()` instead of parsing JSON.
33
34	</td>
35	</tr>
36
37	<tr>
38	<td width="33%">
39
40	### 🔧 100s of Models
41	OpenAI, Anthropic, Mistral, Bedrock, Vertex, Azure. Switch providers per-step or globally.
42
43	</td>
44	<td width="33%">
45
46	### 🔄 Advanced Patterns
47	Parallel execution, branching, loops, sub-agent composition. Enterprise-grade workflow control.
48
49	</td>
50	<td width="33%">
51
52	### 📡 Streaming
53	Stream tokens in real-time as LLMs generate them. Perfect for chat UIs and SSE endpoints.
54
55	</td>
56	</tr>
57
58	<tr>
59	<td width="33%">
60
61	### 🛡️ TypeScript-First
62	Full type safety with IntelliSense. Catch errors before runtime.
63
64	</td>
65	<td width="33%">
66
67	### 📊 Observability
68	OpenTelemetry traces and metrics. Export to Jaeger, Prometheus, DataDog, or any OTLP backend.
69
70	</td>
71	<td width="33%">
72
73	### ⚡ Production-Ready
74	Built-in retries, timeouts, error handling, and connection pooling. Battle-tested at scale.
75
76	</td>
77	</tr>
78	</table>
79
80	[Explore all features →](https://volcano.dev/docs#key-features)
81
82	## Quick Start
83
84	### Installation
85
86	```bash
87	npm install @volcano.dev/agent
88	```
89
90	That's it! Includes MCP support and all common LLM providers (OpenAI, Anthropic, Mistral, Llama, Vertex).
91
92	[View installation guide →](https://volcano.dev/docs#installation)
93
94	### Hello World with Automatic Tool Selection
95
96	```ts
97	import { agent, llmOpenAI, mcp } from "@volcano.dev/agent";
98
99	const llm = llmOpenAI({
100	apiKey: process.env.OPENAI_API_KEY!,
101	model: "gpt-4o-mini"
102	});
103
104	const weather = mcp("http://localhost:8001/mcp");
105	const tasks = mcp("http://localhost:8002/mcp");
106
107	// Agent automatically picks the right tools
108	const results = await agent({ llm })
109	.then({
110	prompt: "What's the weather in Seattle? If it will rain, create a task to bring an umbrella",
111	mcps: [weather, tasks] // LLM chooses which tools to call
112	})
113	.run();
114
115	// Ask questions about what happened
116	const summary = await results.summary(llm);
117	console.log(summary);
118	```
119
120	### Multi-Agent Coordinator
121
122	```ts
123	import { agent, llmOpenAI } from "@volcano.dev/agent";
124
125	const llm = llmOpenAI({ apiKey: process.env.OPENAI_API_KEY! });
126
127	// Define specialized agents
128	const researcher = agent({ llm, name: 'researcher', description: 'Finds facts and data' })
129	.then({ prompt: "Research the topic." })
130	.then({ prompt: "Summarize the research." });
131
132	const writer = agent({ llm, name: 'writer', description: 'Creates content' })
133	.then({ prompt: "Write content." });
134
135	// Coordinator autonomously delegates to specialists
136	const results = await agent({ llm })
137	.then({
138	prompt: "Write a blog post about quantum computing",
139	agents: [researcher, writer] // Coordinator decides when done
140	})
141	.run();
142
143	// Ask what happened
144	const post = await results.ask(llm, "Show me the final blog post");
145	console.log(post);
146	```
147
148	[View more examples →](https://volcano.dev/docs/examples)
149
150	## Documentation
151
152	### 📖 Comprehensive Guides

People who liked this also viewed

6 picks

Type