No API cost AI features

Problem

You want to ship AI features, but:

• hosted API costs scale with usage
• free tiers have hard limits
• you do not want to price or bill users per AI request

Why normal API integration is annoying

Every hosted provider requires:

• an API key tied to your account
• per-token or per-request billing
• rate limits and quota monitoring
• a backend to protect the key

As your app grows, so does your API bill.

How switchboard-ai-sdk solves it

switchboard-ai-sdk moves the cost model to the user side by using AI tools they already have installed. The app developer does not pay for the AI calls; the user runs their own local or authenticated CLI tool.

import { discover, connect } from "switchboard-ai-sdk";

async function generateIdeas(topic: string) {
  const tools = await discover();
  const tool = tools.find((t) => t.available && t.capabilities.includes("chat"));

  if (!tool) {
    throw new Error("Please install Ollama, Codex, Claude Code, or OpenCode.");
  }

  const client = await connect(tool.id);
  const result = await client.chat({
    messages: [{ role: "user", content: `Give me 5 ideas about ${topic}.` }],
  });

  return result.message.content;
}

Free options

• Ollama — run open-source models locally; no API costs at all
• OpenCode free models — opencode/deepseek-v4-flash-free and similar tiers do not require a paid subscription

Limitations

• The user must install and sometimes authenticate the tool
• Free models may have lower quality or rate limits set by the provider
• You still need to handle the case where no tool is available
• Some providers require paid accounts for their best models