Function calling with the Gemini API

Function calling lets you connect models to external tools and APIs. Instead of generating text responses, the model determines when to call specific functions and provides the necessary parameters to execute real-world actions. This allows the model to act as a bridge between natural language and real-world actions and data. Function calling has 3 primary use cases:

  • Augment Knowledge: Access information from external sources like databases, APIs, and knowledge bases.
  • Extend Capabilities: Use external tools to perform computations and extend the limitations of the model, such as using a calculator or creating charts.
  • Take Actions: Interact with external systems using APIs, such as scheduling appointments, creating invoices, sending emails, or controlling smart home devices.

Python

from google import genai

schedule_meeting_function = {
    "type": "function",
    "name": "schedule_meeting",
    "description": "Schedules a meeting with specified attendees at a given time and date.",
    "parameters": {
        "type": "object",
        "properties": {
            "attendees": {"type": "array", "items": {"type": "string"}},
            "date": {"type": "string", "description": "Date (e.g., '2024-07-29')"},
            "time": {"type": "string", "description": "Time (e.g., '15:00')"},
            "topic": {"type": "string", "description": "The meeting topic."},
        },
        "required": ["attendees", "date", "time", "topic"],
    },
}

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    input="Schedule a meeting with Bob and Alice for 03/14/2025 at 10:00 AM about Q3 planning.",
    tools=[{"type": "function", **schedule_meeting_function}],
)

for step in interaction.steps:
    if step.type == "function_call":
        print(f"Function to call: {step.name}")
        print(f"Arguments: {step.arguments}")

JavaScript

import { GoogleGenAI } from '@google/genai';

const client = new GoogleGenAI({});

const scheduleMeetingFunction = {
  type: 'function',
  name: 'schedule_meeting',
  description: 'Schedules a meeting with specified attendees at a given time and date.',
  parameters: {
    type: 'object',
    properties: {
      attendees: { type: 'array', items: { type: 'string' } },
      date: { type: 'string', description: 'Date (e.g., "2024-07-29")' },
      time: { type: 'string', description: 'Time (e.g., "15:00")' },
      topic: { type: 'string', description: 'The meeting topic.' },
    },
    required: ['attendees', 'date', 'time', 'topic'],
  },
};

const interaction = await client.interactions.create({
  model: 'gemini-3-flash-preview',
  input: 'Schedule a meeting with Bob and Alice for 03/27/2025 at 10:00 AM about Q3 planning.',
  tools: [scheduleMeetingFunction],
});

for (const step of interaction.steps) {
  if (step.type === 'function_call') {
    console.log(`Function to call: ${step.name}`);
    console.log(`Arguments: ${JSON.stringify(step.arguments)}`);
  }
}

REST

curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
  -H "x-goog-api-key: $GEMINI_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gemini-3-flash-preview",
    "input": "Schedule a meeting with Bob and Alice for 03/27/2025 at 10:00 AM about Q3 planning.",
    "tools": [{
        "type": "function",
        "name": "schedule_meeting",
        "description": "Schedules a meeting with specified attendees at a given time and date.",
        "parameters": {
          "type": "object",
          "properties": {
            "attendees": {"type": "array", "items": {"type": "string"}},
            "date": {"type": "string"},
            "time": {"type": "string"},
            "topic": {"type": "string"}
          },
          "required": ["attendees", "date", "time", "topic"]
        }
    }]
  }'

How function calling works

function calling overview

Function calling involves a structured interaction between your application, the model, and external functions:

  1. Define Function Declaration: Define the function's name, parameters, and purpose to the model.
  2. Call LLM with function declarations: Send user prompt along with the function declaration(s) to the model.
  3. Execute Function Code (Your Responsibility): The model doesn't execute the function itself. Extract the name and args and execute in your application.
  4. Create User friendly response: Send the result back to the model for a final, user-friendly response.

This process can be repeated over multiple turns. The model supports calling multiple functions in a single turn (parallel function calling) and in sequence (compositional function calling).

Step 1: Define a function declaration

Python

set_light_values_declaration = {
    "type": "function",
    "name": "set_light_values",
    "description": "Sets the brightness and color temperature of a light.",
    "parameters": {
        "type": "object",
        "properties": {
            "brightness": {
                "type": "integer",
                "description": "Light level from 0 to 100",
            },
            "color_temp": {
                "type": "string",
                "enum": ["daylight", "cool", "warm"],
                "description": "Color temperature",
            },
        },
        "required": ["brightness", "color_temp"],
    },
}

def set_light_values(brightness: int, color_temp: str) -> dict:
    """Set the brightness and color temperature of a room light."""
    return {"brightness": brightness, "colorTemperature": color_temp}

JavaScript

const setLightValuesTool = {
  type: 'function',
  name: 'set_light_values',
  description: 'Sets the brightness and color temperature of a light.',
  parameters: {
    type: 'object',
    properties: {
      brightness: { type: 'number', description: 'Light level from 0 to 100' },
      color_temp: { type: 'string', enum: ['daylight', 'cool', 'warm'] },
    },
    required: ['brightness', 'color_temp'],
  },
};

function setLightValues(brightness, color_temp) {
  return { brightness: brightness, colorTemperature: color_temp };
}

Step 2: Call the model with function declarations

Python

from google import genai

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    input="Turn the lights down to a romantic level",
    tools=[set_light_values_declaration],
)

# Find the function call step
fc_step = next(s for s in interaction.steps if s.type == "function_call")
print(fc_step)

JavaScript

import { GoogleGenAI } from '@google/genai';

const client = new GoogleGenAI({});

const interaction = await client.interactions.create({
  model: 'gemini-3-flash-preview',
  input: 'Turn the lights down to a romantic level',
  tools: [setLightValuesTool],
});

// Find the function call step
const fcStep = interaction.steps.find(s => s.type === 'function_call');
console.log(fcStep);

The model returns a function_call step with type, name, and arguments:

type='function_call'
name='set_light_values'
arguments={'color_temp': 'warm', 'brightness': 25}

Step 3: Execute the function

Python

fc_step = next(s for s in interaction.steps if s.type == "function_call")

if fc_step.name == "set_light_values":
    result = set_light_values(**fc_step.arguments)
    print(f"Function execution result: {result}")

JavaScript

const fcStep = interaction.steps.find(s => s.type === 'function_call');

let result;
if (fcStep.name === 'set_light_values') {
  result = setLightValues(fcStep.arguments.brightness, fcStep.arguments.color_temp);
  console.log(`Function execution result: ${JSON.stringify(result)}`);
}

Step 4: Send result back to model

Python

final_interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    input=[
        {
            "type": "function_result",
            "name": fc_step.name,
            "call_id": fc_step.id,
            "result": [{"type": "text", "text": json.dumps(result)}],
        }
    ],
    tools=[set_light_values_declaration],
    previous_interaction_id=interaction.id,
)

print(final_interaction.steps[-1].content[0].text)

JavaScript

const finalInteraction = await client.interactions.create({
  model: 'gemini-3-flash-preview',
  input: [{
    type: 'function_result',
    name: fcStep.name,
    call_id: fcStep.id,
    result: [{ type: 'text', text: JSON.stringify(result) }]
  }],
  tools: [setLightValuesTool],
  previous_interaction_id: interaction.id,
});

console.log(finalInteraction.steps.at(-1).content[0].text);

Function declarations

A function declaration is passed as a tool and includes:

  • type (string): Must be "function" for custom functions.
  • name (string): Unique function name (use underscores or camelCase).
  • description (string): Clear explanation of the function's purpose.
  • parameters (object): Input parameters the function expects.
    • type (string): Overall data type, such as object.
    • properties (object): Individual parameters with type and description.
    • required (array): Mandatory parameter names.

Function calling with thinking models

Gemini 3 and 2.5 series models use an internal "thinking" process that improves function calling. The SDKs automatically handle thought signatures for you.

Parallel function calling

Call multiple functions at once when they are independent:

Python

power_disco_ball = {"type": "function", "name": "power_disco_ball", "description": "Powers the disco ball.",
    "parameters": {"type": "object", "properties": {"power": {"type": "boolean"}}, "required": ["power"]}}
start_music = {"type": "function", "name": "start_music", "description": "Play music.",
    "parameters": {"type": "object", "properties": {"energetic": {"type": "boolean"}, "loud": {"type": "boolean"}}, "required": ["energetic", "loud"]}}
dim_lights = {"type": "function", "name": "dim_lights", "description": "Dim the lights.",
    "parameters": {"type": "object", "properties": {"brightness": {"type": "number"}}, "required": ["brightness"]}}

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    input="Turn this place into a party!",
    tools=[power_disco_ball, start_music, dim_lights],
    generation_config={"tool_choice": "any"},
)

for step in interaction.steps:
    if step.type == "function_call":
        args = ", ".join(f"{key}={val}" for key, val in step.arguments.items())
        print(f"{step.name}({args})")

JavaScript

const powerDiscoBall = { type: 'function', name: 'power_disco_ball', description: 'Powers the disco ball.',
  parameters: { type: 'object', properties: { power: { type: 'boolean' } }, required: ['power'] } };
const startMusic = { type: 'function', name: 'start_music', description: 'Play music.',
  parameters: { type: 'object', properties: { energetic: { type: 'boolean' }, loud: { type: 'boolean' } }, required: ['energetic', 'loud'] } };
const dimLights = { type: 'function', name: 'dim_lights', description: 'Dim the lights.',
  parameters: { type: 'object', properties: { brightness: { type: 'number' } }, required: ['brightness'] } };

const interaction = await client.interactions.create({
  model: 'gemini-3-flash-preview',
  input: 'Turn this place into a party!',
  tools: [powerDiscoBall, startMusic, dimLights],
  generation_config: { tool_choice: 'any' },
});

for (const step of interaction.steps) {
  if (step.type === 'function_call') {
    console.log(`${step.name}(${JSON.stringify(step.arguments)})`);
  }
}

Compositional function calling

Chain multiple function calls together for complex requests (e.g., get location first, then get weather for that location).

Python

get_weather_forecast_declaration = {
    "type": "function",
    "name": "get_weather_forecast",
    "description": "Gets the current weather temperature for a given location.",
    "parameters": {
        "type": "object",
        "properties": {
            "location": {"type": "string", "description": "The location"},
        },
        "required": ["location"],
    },
}

set_thermostat_temperature_declaration = {
    "type": "function",
    "name": "set_thermostat_temperature",
    "description": "Sets the thermostat to a desired temperature.",
    "parameters": {
        "type": "object",
        "properties": {
            "temperature": {
                "type": "integer",
                "description": "The temperature in Celsius",
            },
        },
        "required": ["temperature"],
    },
}

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    input="If it's warmer than 20°C in London, set the thermostat to 20°C, otherwise 18°C.",
    tools=[
        get_weather_forecast_declaration,
        set_thermostat_temperature_declaration,
    ],
)

print(interaction.steps[-1].content[0].text)

Function calling modes

Control how the model uses tools using tool_choice in generation_config:

  • auto (Default): Model decides whether to call a function or respond directly.
  • any: Model is constrained to always predict a function call.
  • none: Model is prohibited from making function calls.

  • validated (Preview): Model ensures function schema adherence.

Python

generation_config = {
    "tool_choice": {
        "allowed_tools": {
            "mode": "any",
            "tools": ["get_current_temperature"]
        }
    }
}

JavaScript

const generation_config = {
  tool_choice: {
    allowed_tools: {
      mode: 'any',
      tools: ['get_current_temperature']
    }
  }
};

REST

curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
  -H "x-goog-api-key: \$GEMINI_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gemini-3-flash-preview",
    "input": "What is the temperature in Boston?",
    "tools": [{
      "type": "function",
      "name": "get_current_temperature",
      "description": "Gets the current temperature for a given location.",
      "parameters": {
        "type": "object",
        "properties": {
          "location": {"type": "string"}
        },
        "required": ["location"]
      }
    }],
    "generation_config": {
      "tool_choice": {
        "allowed_tools": {
          "mode": "any",
          "tools": ["get_current_temperature"]
        }
      }
    }
  }'

Multi-tool use

You can enable multiple tools, combining built-in tools with function calling in the same request. Gemini 3 models can combine built-in tools with function calling out-of-the-box in Interactions. Passing previous_interaction_id automatically circulates the built-in tool context.

Python

from google import genai
import json

client = genai.Client()

get_weather = {
    "type": "function",
    "name": "get_weather",
    "description": "Gets the weather for a requested city.",
    "parameters": {
        "type": "object",
        "properties": {
            "city": {
                "type": "string",
                "description": "The city and state, e.g. Utqiaġvik, Alaska",
            },
        },
        "required": ["city"],
    },
}

tools = [
    {"type": "google_search"},  # Built-in tool
    get_weather                 # Custom tool
]

# Turn 1: Initial request with both tools enabled
interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    input="What is the northernmost city in the United States? What's the weather like there today?",
    tools=tools
)

for step in interaction.steps:
    if step.type == "function_call":
        print(f"Function call: {step.name} (ID: {step.id})")
        # Execute your custom function locally
        result = {"response": "Very cold. 22 degrees Fahrenheit."}
        # Turn 2: Provide the function result back to the model.
        # Passing `previous_interaction_id` automatically circulates the
        # built-in Google Search context from Turn 1
        interaction_2 = client.interactions.create(
            model="gemini-3-flash-preview",
            previous_interaction_id=interaction.id,
            tools=tools,
            input=[{
                "type": "function_result",
                "name": step.name,
                "call_id": step.id,
                "result": [{"type": "text", "text": json.dumps(result)}]
            }]
        )

        print(interaction_2.steps[-1].content[0].text)

JavaScript

import { GoogleGenAI } from '@google/genai';

const client = new GoogleGenAI({});

const weatherTool = {
    type: 'function',
    name: 'get_weather',
    description: 'Gets the weather for a given location.',
    parameters: {
        type: 'object',
        properties: {
            location: { type: 'string', description: 'The city and state, e.g. San Francisco, CA' }
        },
        required: ['location']
    }
};

const tools = [
    {type: 'google_search'}, // Built-in tool
    weatherTool              // Custom tool
];

// Turn 1: Initial request with both tools enabled
let interaction = await client.interactions.create({
    model: 'gemini-3-flash-preview',
    input: "What is the northernmost city in the United States? What's the weather like there today?",
    tools: tools
});

for (const step of interaction.steps) {
    if (step.type === 'function_call') {
        console.log(`Function call: ${step.name} (ID: ${step.id})`);
        // Execute your custom function locally
        const result = {response: "Very cold. 22 degrees Fahrenheit."};
        // Turn 2: Provide the function result back to the model.
        const interaction_2 = await client.interactions.create({
            model: 'gemini-3-flash-preview',
            previous_interaction_id: interaction.id,
            tools: tools,
            input: [{
                type: 'function_result',
                name: step.name,
                call_id: step.id,
                result: [{ type: 'text', text: JSON.stringify(result) }]
            }]
        });

        console.log(interaction_2.steps.at(-1).content[0].text);
    }
}

Multimodal function responses

For Gemini 3 series models, you can include multimodal content in the function response parts that you send to the model. The model can process this multimodal content in its next turn to produce a more informed response.

To include multimodal data in a function response, include it as one or more content blocks in the result field of the function_result step. Each content block must specify its type (e.g., "text", "image").

The following example shows how to send a function response containing image data back to the model in an interaction:

Python

import base64
from google import genai
import requests

client = genai.Client()

# Find the function call step
tool_call = next(s for s in interaction.steps if s.type == "function_call")

# Execute your tool to get image bytes
image_path = "https://goo.gle/instrument-img"
image_bytes = requests.get(image_path).content

base64_image_data = base64.b64encode(image_bytes).decode("utf-8")

final_interaction = client.interactions.create(
    model="gemini-3-flash-preview",
    previous_interaction_id=interaction.id,
    input=[
        {
            "type": "function_result",
            "name": tool_call.name,
            "call_id": tool_call.id,
            "result": [
                {"type": "text", "text": "instrument.jpg"},
                {
                    "type": "image",
                    "mime_type": "image/jpeg",
                    "data": base64_image_data,
                },
            ],
        }
    ],
)

print(final_interaction.steps[-1].content[0].text)

JavaScript

import { GoogleGenAI } from "@google/genai";

const ai = new GoogleGenAI({});

// Find the function call step
const toolCall = interaction.steps.find(s => s.type === 'function_call');

// Execute your tool to get image bytes and convert to base64
// (Implementation depends on your environment)
const base64ImageData = "BASE64_IMAGE_DATA";

const finalInteraction = await ai.interactions.create({
    model: 'gemini-3-flash-preview',
    previousInteractionId: interaction.id,
    input: [{
        type: 'function_result',
        name: toolCall.name,
        call_id: toolCall.id,
        result: [
            { type: 'text', text: 'instrument.jpg' },
            {
                type: 'image',
                mime_type: 'image/jpeg',
                data: base64ImageData,
            }
        ]
    }]
});

console.log(finalInteraction.steps.at(-1).content[0].text);

REST

curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
  -H "x-goog-api-key: \$GEMINI_API_KEY" \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gemini-3-flash-preview",
    "previous_interaction_id": "INTERACTION_ID",
    "input": [
      {
        "type": "function_result",
        "name": "get_image",
        "call_id": "call_123",
        "result": [
          {"type": "text", "text": "instrument.jpg"},
          {
            "type": "image",
            "mime_type": "image/jpeg",
            "data": "BASE64_IMAGE_DATA"
          }
        ]
      }
    ]
  }'

Function calling with Structured output

For Gemini 3 series models, combine function calling with structured output for consistently formatted responses.

Remote MCP (Model Context Protocol)

Interactions API supports connecting to remote MCP servers to give the model access to external tools and services. You provide the server name and url in the tools configuration.

When using Remote MCP, be aware of the following constraints:

  • Server types: Remote MCP only works with Streamable HTTP servers. SSE (Server-Sent Events) servers are not supported.
  • Model support: Remote MCP does not work with Gemini 3 models at this time. Support for Gemini 3 is coming soon.
  • Naming: MCP server names should not include the - character. Use snake_case server names instead.
Field Type Required Description
type string Yes Must be "mcp_server".
name string No A display name for the MCP server.
url string No The full URL for the MCP server endpoint.
headers object No Key-value pairs sent as HTTP headers with every request to the server (for example, authentication tokens).
allowed_tools array No Restrict which tools from the server the agent may call.

Example

Python

from google import genai

client = genai.Client()

interaction = client.interactions.create(
    model="gemini-2.5-flash",
    input="Check the status of my last server deployment.",
    tools=[
        {
            "type": "mcp_server",
            "name": "Deployment Tracker",
            "url": "https://mcp.example.com/mcp",
            "headers": {"Authorization": "Bearer my-token"},
        }
    ]
)

JavaScript

import { GoogleGenAI } from '@google/genai';

const client = new GoogleGenAI({});

const interaction = await client.interactions.create({
    model: 'gemini-2.5-flash',
    input: 'Check the status of my last server deployment.',
    tools: [
        {
            type: 'mcp_server',
            name: 'Deployment Tracker',
            url: 'https://mcp.example.com/mcp',
            headers: { Authorization: 'Bearer my-token' }
        }
    ]
});

REST

curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-d '{
    "model": "gemini-2.5-flash",
    "input": "Check the status of my last server deployment.",
    "tools": [
        {
            "type": "mcp_server",
            "name": "Deployment Tracker",
            "url": "https://mcp.example.com/mcp",
            "headers": {"Authorization": "Bearer my-token"}
        }
    ]
}'

Stream tool calls

When using tools with streaming, the model generates function calls as a sequence of step.delta events on the stream. Tool arguments can be streamed as partial arguments using arguments. You must aggregate these deltas to reconstruct the complete tool calls before executing them.

Python

import json
from google import genai

client = genai.Client()

weather_tool = {
    "type": "function",
    "name": "get_weather",
    "description": "Gets the weather for a given location.",
    "parameters": {
        "type": "object",
        "properties": {
            "location": {"type": "string", "description": "The city and state"}
        },
        "required": ["location"]
    }
}

stream = client.interactions.create(
    model="gemini-3-flash-preview",
    input="What is the weather in Paris?",
    tools=[weather_tool],
    stream=True
)

current_calls = {}
tool_calls = []

for event in stream:
    if event.event_type == "step.start":
        if event.step.type == "function_call":
            current_calls[event.index] = {
                "id": event.step.id,
                "name": event.step.name,
                "arguments": ""
            }
            # Handle arguments provided in step.start
            if hasattr(event.step, "arguments") and event.step.arguments:
                if isinstance(event.step.arguments, dict):
                    current_calls[event.index]["arguments"] = json.dumps(event.step.arguments)
                else:
                    current_calls[event.index]["arguments"] = event.step.arguments
    elif event.event_type == "step.delta":
        if event.delta.type == "arguments":
            if event.index in current_calls:
                current_calls[event.index]["arguments"] += event.delta.partial_arguments
        elif event.delta.type == "text":
            print(event.delta.text, end="", flush=True)

    elif event.event_type == "interaction.completed":
        for index, call in current_calls.items():
            args = call["arguments"]
            if args:
                args = json.loads(args)
            else:
                args = {}

            tool_calls.append({
                "type": "function_call",
                "id": call["id"],
                "name": call["name"],
                "arguments": args
            })

        print(f"\nFinal tool calls ready to execute:")
        print(json.dumps(tool_calls, indent=2))

JavaScript

import { GoogleGenAI } from '@google/genai';

const client = new GoogleGenAI({});

const weatherTool = {
    type: 'function',
    name: 'get_weather',
    description: 'Gets the weather for a given location.',
    parameters: {
        type: 'object',
        properties: {
            location: { type: 'string', description: 'The city and state' }
        },
        required: ['location']
    }
};

const stream = await client.interactions.create({
    model: 'gemini-3-flash-preview',
    input: 'What is the weather in Paris?',
    tools: [weatherTool],
    stream: true,
});

const currentCalls = new Map();
let toolCalls = [];

for await (const event of stream) {
    const evType = event.event_type;
    if (evType === 'step.start') {
        if (event.step.type === 'function_call') {
            currentCalls.set(event.index, {
                id: event.step.id,
                name: event.step.name,
                arguments: ''
            });
            // Handle arguments provided in step.start
            if (event.step.arguments) {
                if (typeof event.step.arguments === 'object') {
                    currentCalls.get(event.index).arguments = JSON.stringify(event.step.arguments);
                } else {
                    currentCalls.get(event.index).arguments = event.step.arguments;
                }
            }
        }
    } else if (evType === 'step.delta') {
        if (event.delta.type === 'arguments') {
            if (currentCalls.has(event.index)) {
                currentCalls.get(event.index).arguments += event.delta.partial_arguments;
            }
        } else if (event.delta.type === 'text') {
            process.stdout.write(event.delta.text);
        }
    } else if (evType === 'interaction.completed' || evType === 'interaction.complete') {
        toolCalls = Array.from(currentCalls.values()).map(call => ({
            type: 'function_call',
            id: call.id,
            name: call.name,
            arguments: call.arguments ? JSON.parse(call.arguments) : {}
        }));
        console.log('\nFinal tool calls ready to execute:');
        console.log(JSON.stringify(toolCalls, null, 2));
    }
}

REST

curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions?alt=sse" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-d '{
    "model": "gemini-3-flash-preview",
    "input": "What is the weather in Paris?",
    "tools": [{
        "type": "function",
        "name": "get_weather",
        "description": "Gets the weather for a given location.",
        "parameters": {
            "type": "object",
            "properties": {
                "location": {"type": "string", "description": "The city and state"}
            },
            "required": ["location"]
        }
    }],
    "stream": true
}'

Best practices

  • Function and Parameter Descriptions: Be clear and specific.
  • Naming: Use descriptive names without spaces or special characters.
  • Strong Typing: Use specific types (integer, string, enum).
  • Tool Selection: Keep active set to 10-20 tools maximum.
  • Prompt Engineering: Provide context and instructions.
  • Validation: Validate function calls before executing.
  • Error Handling: Implement robust error handling.
  • Security: Use appropriate authentication for external APIs.

Notes and limitations

  • Only a subset of the OpenAPI schema is supported.
  • For any mode, the API may reject very large or deeply nested schemas.
  • Supported parameter types in Python are limited.