pages/antigravity-agent.txt

route: /gemini-api/docs/antigravity-agent
title: Antigravity Agent
description: Use the Antigravity managed agent to perform multi-step tasks that require reasoning, file operations, and tool use.

The Antigravity agent is a general-purpose managed agent on the Gemini API. A single API call gives you an agent that reasons, executes code, manages files, and browses the web inside your own secure Linux sandbox, hosted by Google.
It is powered by Gemini 3.5 Flash and uses the same harness as the Antigravity IDE. Available through the Interactions API and Google AI Studio.
Python
from google import genai
client = genai.Client()
interaction = client.interactions.create(
agent="antigravity-preview-05-2026",
input="Read Hacker News, summarize the top 10 stories, and save the results as a PDF.",
environment="remote",
)
print(interaction.output_text)
JavaScript
import { GoogleGenAI } from "@google/genai";
const client = new GoogleGenAI({});
const interaction = await client.interactions.create({
agent: "antigravity-preview-05-2026",
input: "Read Hacker News, summarize the top 10 stories, and save the results as a PDF.",
environment: "remote",
}, { timeout: 300000 });
console.log(interaction.output_text);
REST
curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-d '{
"agent": "antigravity-preview-05-2026",
"input": "Read Hacker News, summarize the top 10 stories, and save the results as a PDF.",
"environment": "remote"
}'
Capabilities
Each call can provision a Linux sandbox and starts a tool-use loop. The agent plans, acts, observes results, and repeats until the task is done.
Code execution: Run Bash, Python, and Node.js commands. Install packages, run tests, build apps.
File management: Read, write, edit, search, and list files in the sandbox. Files persist across interactions.
Web access: Google Search and URL fetching for data.
Context compaction: Automatic context compaction (triggered at ~135k tokens) to support long-running, multi-turn sessions without losing context or hitting token limits.
See the Quickstart for multi-turn usage and streaming.
Supported tools
By default, the agent has access to code_execution, google_search, and url_context. Filesystem tools are enabled automatically when you specify the environment parameter. You can also define custom functions to connect the agent to your own APIs and tools. You only need to specify the tools parameter when customizing or restricting the default set, or when adding custom functions.
Tool
Type value
Description
Code Execution
code_execution
Run shell commands (bash, Python, Node) with stdout/stderr capture.
Google Search
google_search
Search the public web.
URL Context
url_context
Fetch and read web pages.
Filesystem
(enabled via environment)
Read, write, edit, search, and list files in the sandbox. No separate tool type; enabled automatically when environment is set.
Custom Functions
function
Define custom functions that the agent can request to execute. See Function calling.
Remote MCP Server
mcp_server
Register external Model Context Protocol (MCP) servers as tools. See MCP servers.
To limit the agent to specific tools, pass only the ones you need:
Python
from google import genai
client = genai.Client()
interaction = client.interactions.create(
agent="antigravity-preview-05-2026",
input="Search for the latest AI research papers on reasoning and summarize them.",
environment="remote",
tools=[
{"type": "google_search"},
{"type": "url_context"},
],
)
print(interaction.output_text)
JavaScript
import { GoogleGenAI } from "@google/genai";
const client = new GoogleGenAI({});
const interaction = await client.interactions.create({
agent: "antigravity-preview-05-2026",
input: "Search for the latest AI research papers on reasoning and summarize them.",
environment: "remote",
tools: [
{ type: "google_search" },
{ type: "url_context" },
],
}, { timeout: 300000 });
console.log(interaction.output_text);
REST
curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-d '{
"agent": "antigravity-preview-05-2026",
"input": "Search for the latest AI research papers on reasoning and summarize them.",
"environment": "remote",
"tools": [
{"type": "google_search"},
{"type": "url_context"}
]
}'
Multimodal Input
The Antigravity agent supports multimodal inputs. Currently, only text and image inputs are supported. Images must be supplied as inline base64-encoded strings (data).
Python
import base64
from google import genai
client = genai.Client()
with open("path/to/chart.png", "rb") as f:
image_bytes = f.read()
interaction_inline = client.interactions.create(
agent="antigravity-preview-05-2026",
input=[
{"type": "text", "text": "Analyze this chart and summarize the trends."},
{
"type": "image",
"data": base64.b64encode(image_bytes).decode("utf-8"),
"mime_type": "image/png",
},
],
environment="remote",
)
JavaScript
import { GoogleGenAI } from "@google/genai";
import * as fs from "node:fs";
const client = new GoogleGenAI({});
const base64Image = fs.readFileSync("path/to/chart.png", { encoding: "base64" });
const interactionInline = await client.interactions.create({
agent: "antigravity-preview-05-2026",
input: [
{ type: "text", text: "Analyze this chart and summarize the trends." },
{
type: "image",
data: base64Image,
mime_type: "image/png",
},
],
environment: "remote",
}, { timeout: 300000 });
REST
BASE64_IMAGE=$(base64 -w0 /path/to/chart.png)
curl -X POST "https://generativelanguage.googleapis.com/v1beta/interactions" \
-H "Content-Type: application/json" \
-H "x-goog-api-key: $GEMINI_API_KEY" \
-d "{
\"agent\": \"antigravity-preview-05-2026\",
\"input\": [
{\"type\": \"text\", \"text\": \"Analyze this chart and summarize the trends.\"},
{
\"type\": \"image\",
\"mime_type\": \"image/png\",
\"data\": \"$BASE64_IMAGE\"
}
],
\"environment\": \"remote\"
}"
Function calling
Function calling allows you to connect the Antigravity agent to external APIs and databases by defining custom tools the agent can invoke. For general concepts, see Function calling with the Gemini API.
The following example demonstrates a 2-turn interaction. The agent first requests a custom get_weather function call, and the client executes it and returns the result in the second turn.
Python
from google import genai
client = genai.Client()
# 1. Define the custom function
get_weather_tool = {
"type": "function",
"name": "get_weather",
"description": "Gets the current weather for a given location.",
"parameters": {
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "The city and country, e.g. San Francisco, USA",
}
},
"required": ["location"],
},
}
# 2. Call the agent with the custom tool (Turn 1)
interaction = client.interactions.create(
agent="antigravity-preview-05-2026",
input="What is the weather in Tokyo?",
environment="remote",
tools=[
{"type": "code_execution"}, # Enable default code execution
get_weather_tool, # Add custom function
],
)
# Check if the agent requested a function call
if interaction.status == "requires_action":
# Find function calls that do not have a matching function result.
# Filesystem tools (like write_file) are also represented as function calls
# but are executed automatically by the environment.
executed_calls = {step.call_id for step in interaction.steps if step.type == "function_result"}
pending_calls = [step for step in interaction.steps if step.type == "function_call" and step.id not in executed_calls]
if pending_calls:
fc_step = pending_calls[0]
print(f"Function to call: {fc_step.name} (ID: {fc_step.id})")
print(f"Arguments: {fc_step.arguments}")
# 3. Execute the function locally (simulated get_weather()) and send the result back (Turn 2)
function_result = {
"temperature": 23,
"unit": "celsius"
}
final_interaction = client.interactions.create(
agent="antigravity-preview-05-2026",
previous_int
…