Build your own
Cursor
with Zypher Agent

Just a few lines of code to create powerful AI agents. Connect any MCP server, choose your LLM provider, and start building immediately.

Get Started
main.ts
const agent = new ZypherAgent(
new AnthropicModelProvider({
apiKey: getRequiredEnv("ANTHROPIC_API_KEY"),
})
);
await agent.mcpServerManager.registerServer({
id: "firecrawl",
type: "command",
command: {
command: "npx",
args: ["-y", "firecrawl-mcp"],
env: {
FIRECRAWL_API_KEY: getRequiredEnv("FIRECRAWL_API_KEY"),
},
},
});
await agent.init();
const event$ = agent.runTask(
`Find latest AI news`,
"claude-sonnet-4-20250514"
);
for await (const event of eachValueFrom(event$)) {
console.log(event);
}

Everything you need to build AI agents

🤖

Interactive CLI

Fast prototyping with an intuitive command-line interface for building and testing agents

🛠️

Tool Calling & MCP

Built-in tool system with Model Context Protocol support for seamless integrations

📝

Git Checkpoints

Track and revert changes with a git-based checkpoint system for reliable development

Trusted by developers building the future of AI