1mcp FAQs

Question 1

Can I integrate 1mcp with my existing AI development workflow?

Accepted Answer

Absolutely. 1mcp offers native integration with AI SDK through the `@onemcp/ai-sdk` package, which automatically converts your AI SDK tools into a sandboxed MCP surface. It also supports browser-side execution to offload compute.

Question 2

What is 1mcp and its primary purpose?

Accepted Answer

1mcp is a tool that empowers AI agents to safely write and execute code within a sandboxed WebAssembly (WASM) environment. Its primary purpose is to significantly reduce token usage for chained tool calls by running them as a single execution.

Question 3

Is code execution with 1mcp secure?

Accepted Answer

Yes, 1mcp provides a highly secure environment. It executes code in a sandboxed WASM runtime with configurable, policy-enforced limits on network access, filesystem operations, and runtime resources (like memory and timeout), ensuring safe operations.

Question 4

What are the key advantages of 1mcp over general cloud sandboxes?

Accepted Answer

1mcp excels with chained execution for massive token reduction, supports standard TypeScript/JavaScript with npm dependencies (no provider-specific SDKs), provides fine-grained session-scoped security policies, and allows browser execution to offload server compute, offering specialized benefits beyond generic cloud sandboxes.

Question 5

How does 1mcp achieve significant token reduction?

Accepted Answer

1mcp reduces token usage by allowing AI agents to compose multiple Model Context Protocol (MCP) tool calls into a single JavaScript/TypeScript code execution. Instead of making individual API calls, the agent's code chains them, leading to up to a 96% reduction in tokens compared to separate prompts for each action.

1mcp

1mcp

主要功能

使用案例

主要功能

使用案例