Tools

Function Calling

Define custom tools that the model can invoke during a conversation. The model requests the call, you execute it locally, and return the result. This enables integration with databases, APIs, and any external system.

WARNING

With streaming, the function call is returned in whole in a single chunk, not streamed across chunks.

1. Define tools with a name, description, and JSON schema for parameters
2. Include tools in your request
3. Model returns a tool_call when it needs external data
4. Execute the function locally and return the result
5. Model continues with your result

Quick Start

curl https://api.x.ai/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $XAI_API_KEY" \
  -d '{
  "model": "grok-4.5",
  "input": [
    {"role": "user", "content": "What is the temperature in San Francisco?"}
  ],
  "tools": [
    {
      "type": "function",
      "name": "get_temperature",
      "description": "Get current temperature for a location",
      "parameters": {
        "type": "object",
        "properties": {
          "location": {"type": "string", "description": "City name"},
          "unit": {"type": "string", "enum": ["celsius", "fahrenheit"], "default": "fahrenheit"}
        },
        "required": ["location"]
      }
    }
  ]
}'

import os
import json

from xai_sdk import Client
from xai_sdk.chat import user, tool, tool_result

client = Client(api_key=os.getenv("XAI_API_KEY"))

# Define tools
tools = [
    tool(
        name="get_temperature",
        description="Get current temperature for a location",
        parameters={
            "type": "object",
            "properties": {
                "location": {"type": "string", "description": "City name"},
                "unit": {"type": "string", "enum": ["celsius", "fahrenheit"], "default": "fahrenheit"}
            },
            "required": ["location"]
        },
    ),
]

chat = client.chat.create(
    model="grok-4.5",
    tools=tools,
)
chat.append(user("What is the temperature in San Francisco?"))
response = chat.sample()

# Handle tool calls
if response.tool_calls:
    chat.append(response)
    for tc in response.tool_calls:
        args = json.loads(tc.function.arguments)
        # Execute your function
        result = {"location": args["location"], "temperature": 59, "unit": args.get("unit", "fahrenheit")}
        chat.append(tool_result(json.dumps(result)))

    response = chat.sample()

print(response.content)

import os
import json
from openai import OpenAI

client = OpenAI(
    api_key=os.getenv("XAI_API_KEY"),
    base_url="https://api.x.ai/v1",
)

tools = [
    {
        "type": "function",
        "name": "get_temperature",
        "description": "Get current temperature for a location",
        "parameters": {
            "type": "object",
            "properties": {
                "location": {"type": "string", "description": "City name"},
                "unit": {"type": "string", "enum": ["celsius", "fahrenheit"], "default": "fahrenheit"}
            },
            "required": ["location"]
        },
    },
]

response = client.responses.create(
    model="grok-4.5",
    input=[{"role": "user", "content": "What is the temperature in San Francisco?"}],
    tools=tools,
)

# Handle function calls
for item in response.output:
    if item.type == "function_call":
        args = json.loads(item.arguments)
        result = {"location": args["location"], "temperature": 59, "unit": args.get("unit", "fahrenheit")}

        response = client.responses.create(
            model="grok-4.5",
            input=[{"type": "function_call_output", "call_id": item.call_id, "output": json.dumps(result)}],
            tools=tools,
            previous_response_id=response.id,
        )

for item in response.output:
    if item.type == "message":
        print(item.content[0].text)

import { xai } from '@ai-sdk/xai';
import { streamText, tool, stepCountIs } from 'ai';
import { z } from 'zod';

const result = streamText({
  model: xai.responses('grok-4.5'),
  tools: {
    getTemperature: tool({
      description: 'Get current temperature for a location',
      inputSchema: z.object({
        location: z.string().describe('City name'),
        unit: z.enum(['celsius', 'fahrenheit']).default('fahrenheit'),
      }),
      execute: async ({ location, unit }) => ({
        location,
        temperature: unit === 'fahrenheit' ? 59 : 15,
        unit,
      }),
    }),
  },
  stopWhen: stepCountIs(5),
  prompt: 'What is the temperature in San Francisco?',
});

for await (const chunk of result.fullStream) {
  if (chunk.type === 'text-delta') {
    process.stdout.write(chunk.text);
  }
}

Defining Tools with Pydantic

Use Pydantic models for type-safe parameter schemas:

from typing import Literal
from pydantic import BaseModel, Field
from xai_sdk.chat import tool

class TemperatureRequest(BaseModel):
    location: str = Field(description="City and state, e.g. San Francisco, CA")
    unit: Literal["celsius", "fahrenheit"] = Field("fahrenheit", description="Temperature unit")

class CeilingRequest(BaseModel):
    location: str = Field(description="City and state, e.g. San Francisco, CA")

# Generate JSON schema from Pydantic models
tools = [
    tool(
        name="get_temperature",
        description="Get current temperature for a location",
        parameters=TemperatureRequest.model_json_schema(),
    ),
    tool(
        name="get_ceiling",
        description="Get current cloud ceiling for a location",
        parameters=CeilingRequest.model_json_schema(),
    ),
]

from typing import Literal
from pydantic import BaseModel, Field

class TemperatureRequest(BaseModel):
    location: str = Field(description="City and state, e.g. San Francisco, CA")
    unit: Literal["celsius", "fahrenheit"] = Field("fahrenheit", description="Temperature unit")

class CeilingRequest(BaseModel):
    location: str = Field(description="City and state, e.g. San Francisco, CA")

tools = [
    {
        "type": "function",
        "name": "get_temperature",
        "description": "Get current temperature for a location",
        "parameters": TemperatureRequest.model_json_schema(),
    },
    {
        "type": "function",
        "name": "get_ceiling",
        "description": "Get current cloud ceiling for a location",
        "parameters": CeilingRequest.model_json_schema(),
    },
]

Handling Tool Calls

When the model wants to use your tool, execute the function and return the result:

import json

def get_temperature(location: str, unit: str = "fahrenheit") -> dict:
    # In production, call a real weather API
    temp = 59 if unit == "fahrenheit" else 15
    return {"location": location, "temperature": temp, "unit": unit}

def get_ceiling(location: str) -> dict:
    return {"location": location, "ceiling": 15000, "unit": "ft"}

tools_map = {
    "get_temperature": get_temperature,
    "get_ceiling": get_ceiling,
}

chat.append(user("What's the weather in Denver?"))
response = chat.sample()

# Process tool calls
if response.tool_calls:
    chat.append(response)

    for tool_call in response.tool_calls:
        name = tool_call.function.name
        args = json.loads(tool_call.function.arguments)

        result = tools_map[name](**args)
        chat.append(tool_result(json.dumps(result)))

    response = chat.sample()

print(response.content)

import json

def get_temperature(location: str, unit: str = "fahrenheit") -> dict:
    temp = 59 if unit == "fahrenheit" else 15
    return {"location": location, "temperature": temp, "unit": unit}

tools_map = {"get_temperature": get_temperature}

# Process function calls
for item in response.output:
    if item.type == "function_call":
        name = item.name
        args = json.loads(item.arguments)

        if name not in tools_map:
            output = json.dumps({"error": f"Unknown function: {name}"})
        else:
            output = json.dumps(tools_map[name](**args))

        response = client.responses.create(
            model="grok-4.5",
            input=[{"type": "function_call_output", "call_id": item.call_id, "output": output}],
            tools=tools,
            previous_response_id=response.id,
        )

for item in response.output:
    if item.type == "message":
        print(item.content[0].text)

Combining with Built-in Tools

Function calling works alongside built-in agentic tools. The model can use web search, then call your custom function:

from xai_sdk.chat import tool
from xai_sdk.tools import web_search, x_search

tools = [
    web_search(),                    # Built-in: runs on xAI servers
    x_search(),                      # Built-in: runs on xAI servers
    tool(                            # Custom: runs on your side
        name="save_to_database",
        description="Save research results to the database",
        parameters={
            "type": "object",
            "properties": {
                "data": {"type": "string", "description": "Data to save"}
            },
            "required": ["data"]
        },
    ),
]

chat = client.chat.create(
    model="grok-4.5",
    tools=tools,
)

tools = [
    {"type": "web_search"},          # Built-in
    {"type": "x_search"},            # Built-in
    {                                # Custom
        "type": "function",
        "name": "save_to_database",
        "description": "Save research results to the database",
        "parameters": {
            "type": "object",
            "properties": {
                "data": {"type": "string", "description": "Data to save"}
            },
            "required": ["data"]
        },
    },
]

When mixing tools:

• Built-in tools execute automatically on xAI servers
• Custom tools pause execution and return to you for handling

See Advanced Usage for complete examples with tool loops.

Tool Choice

Control when the model uses tools:

Value	Behavior
"auto"	Model decides whether to call a tool (default)
"required"	Model must call at least one tool
"none"	Disable tool calling
{"type": "function", "function": {"name": "..."}}	Force a specific tool

Parallel Function Calling

By default, parallel function calling is enabled — the model can request multiple tool calls in a single response. Process all of them before continuing:

# response.tool_calls may contain multiple calls
for tool_call in response.tool_calls:
    result = tools_map[tool_call.function.name](**json.loads(tool_call.function.arguments))
    # Append each result...

Disable with parallel_tool_calls: false in your request.

Tool Schema Reference

Field	Required	Description
name	Yes	Unique identifier (max 200 tools per request)
description	Yes	What the tool does — helps the model decide when to use it
parameters	Yes	JSON Schema defining function inputs

Parameter Schema

{
  "type": "object",
  "properties": {
    "location": {
      "type": "string",
      "description": "City name"
    },
    "unit": {
      "type": "string",
      "enum": ["celsius", "fahrenheit"],
      "default": "celsius"
    }
  },
  "required": ["location"]
}

The root of a parameters schema must be an object ("type": "object"); nest any other types inside properties. A root anyOf or oneOf also works when every branch is itself an object, letting you define a tool that accepts one of several object variants:

{
  "oneOf": [
    {
      "type": "object",
      "properties": {
        "kind": { "const": "email" },
        "address": { "type": "string" }
      },
      "required": ["kind", "address"]
    },
    {
      "type": "object",
      "properties": {
        "kind": { "const": "sms" },
        "phone": { "type": "string" }
      },
      "required": ["kind", "phone"]
    }
  ]
}

WARNING

A tool whose parameters root is neither an object nor a union of objects (for example, a scalar, an array, or an anyOf/oneOf with a non-object branch) cannot be compiled into a tool-call grammar and is rejected with a 400 error that names the tool.

Complete Vercel AI SDK Example

The Vercel AI SDK handles tool definition, execution, and the request/response loop automatically:

import { xai } from '@ai-sdk/xai';
import { streamText, tool, stepCountIs } from 'ai';
import { z } from 'zod';

const result = streamText({
  model: xai.responses('grok-4.5'),
  tools: {
    getCurrentTemperature: tool({
      description: 'Get current temperature for a location',
      inputSchema: z.object({
        location: z.string().describe('City and state, e.g. San Francisco, CA'),
        unit: z.enum(['celsius', 'fahrenheit']).default('fahrenheit'),
      }),
      execute: async ({ location, unit }) => ({
        location,
        temperature: unit === 'fahrenheit' ? 59 : 15,
        unit,
      }),
    }),
    getCurrentCeiling: tool({
      description: 'Get current cloud ceiling for a location',
      inputSchema: z.object({
        location: z.string().describe('City and state'),
      }),
      execute: async ({ location }) => ({
        location,
        ceiling: 15000,
        ceiling_type: 'broken',
        unit: 'ft',
      }),
    }),
  },
  stopWhen: stepCountIs(5),
  prompt: "What's the temperature and cloud ceiling in San Francisco?",
});

for await (const chunk of result.fullStream) {
  switch (chunk.type) {
    case 'text-delta':
      process.stdout.write(chunk.text);
      break;
    case 'tool-call':
      console.log(`Tool call: ${chunk.toolName}`, chunk.input);
      break;
    case 'tool-result':
      console.log(`Tool result: ${chunk.toolName}`, chunk.output);
      break;
  }
}