Tool Calling

Give models structured tools and execute calls in your application loop

TheRouter.ai normalizes tool calling across providers. You pass OpenAI-compatible tool definitions, the model returns `tool_calls`, then your app executes the tool and sends results back in a follow-up completion.

Tool definition and choices

tool-call-request.json

{
  "model": "openai/gpt-4.1",
  "messages": [{ "role": "user", "content": "Check weather in Tokyo and Seoul." }],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get current weather by city",
        "parameters": {
          "type": "object",
          "properties": { "city": { "type": "string" } },
          "required": ["city"]
        }
      }
    }
  ],
  "tool_choice": "auto",
  "parallel_tool_calls": true
}

Execution loop

TypeScript

const first = await client.chat.completions.create(request);
const message = first.choices[0]?.message;

if (message?.tool_calls?.length) {
  const toolMessages = await Promise.all(
    message.tool_calls.map(async (call) => {
      const args = JSON.parse(call.function.arguments || "{}");
      const result = await runTool(call.function.name, args);
      return {
        role: "tool" as const,
        tool_call_id: call.id,
        content: JSON.stringify(result),
      };
    })
  );

  const second = await client.chat.completions.create({
    ...request,
    messages: [...request.messages, message, ...toolMessages],
  });

  console.log(second.choices[0]?.message?.content);
}

Support matrix guidance

Use the models catalog and filter for `tools` support before enabling forced tool-choice behavior. Some models support tool calls but not parallel calls.

Key parameters

Use tool_choice: "auto" to let the model decide, "none" to disable tool use, or force a specific function by name when deterministic behavior is required.