antigravity-sdk-python

Name: antigravity-sdk-python
Author: google-antigravity

263kinstalls/mo · pypi2.6kstars961forks

How to install antigravity-sdk-python?

google-antigravity/antigravity-sdk-python

$git clone https://github.com/google-antigravity/antigravity-sdk-python

Installs into the current project.

Summary

The Google Antigravity SDK is a Python SDK for building AI agents powered by Antigravity and Gemini. It provides a secure, scalable, and stateful infrastructure layer that abstracts the agentic loop, letting you focus on what your agent *does* rather than how it runs.

Files · 1

View on GitHub

README.md

1	# Google Antigravity SDK
2
3	The Google Antigravity SDK is a Python SDK for building AI agents powered by
4	Antigravity and Gemini. It provides a secure, scalable, and stateful
5	infrastructure layer that abstracts the agentic loop, letting you focus on what
6	your agent does rather than how it runs.
7
8	## Installation
9
10	```sh
11	pip install google-antigravity
12	```
13
14	> [!IMPORTANT]
15	> The Google Antigravity SDK relies on a compiled runtime binary that is
16	> included in the platform-specific wheels published to
17	> [PyPI](https://pypi.org/project/google-antigravity/). **Cloning this
18	> repository alone is not sufficient to run the SDK.** Always install from
19	> PyPI with `pip install google-antigravity` to obtain the binary.
20
21	## Quickstart
22
23	Get started by running one of the [`examples/`](examples/), such as the
24	`hello_world` example with:
25
26	```sh
27	export GEMINI_API_KEY="your_api_key_here"
28	python ./examples/getting_started/hello_world.py
29	```
30	## Gemini Enterprise Agent Platform (formerly Vertex AI)
31
32	To use the SDK with Gemini Enterprise Agent Platform (formerly Vertex AI),
33	configure `LocalAgentConfig` with `vertex=True` and specify your GCP `project`
34	and `location`.
35
36	By default, the SDK uses Application Default Credentials (ADC) for
37	authentication.
38
39	```python
40	from google.antigravity import Agent, LocalAgentConfig
41
42	config = LocalAgentConfig(
43	vertex=True,
44	project="your-gcp-project",
45	location="us-central1",
46	)
47
48	async with Agent(config) as agent:
49	response = await agent.chat("Hello!")
50	print(await response.text())
51	```
52
53	Alternatively, you can leave these fields unset in `LocalAgentConfig` and
54	export the environment variables instead:
55
56	```sh
57	# Either GOOGLE_GENAI_USE_VERTEXAI or GOOGLE_GENAI_USE_ENTERPRISE enable Vertex.
58	export GOOGLE_GENAI_USE_VERTEXAI=True
59	export GOOGLE_CLOUD_PROJECT="your-gcp-project"
60	export GOOGLE_CLOUD_LOCATION="us-central1"
61	```
62
63	Explicit kwargs always take precedence over env vars.
64
65	Ensure you have authenticated locally before running the agent:
66
67	```sh
68	gcloud auth application-default login
69	```
70
71	## Concepts
72
73	### Simple Agent
74
75	The `Agent` class is the easiest way to get started. It manages the full
76	lifecycle — binary discovery, tool wiring, hook registration, and policy
77	defaults — behind a single async context manager.
78
79	The `system_instructions` parameter is optional.
80
81	```python
82	import asyncio
83	from google.antigravity import Agent, LocalAgentConfig
84
85	async def main():
86	config = LocalAgentConfig(
87	system_instructions="You are an expert assistant for codebase navigation.",
88	# api_key="your_api_key_here",
89	)
90	async with Agent(config) as agent:
91	response = await agent.chat("What files are in the current directory?")
92	print(await response.text())
93
94	async def run():
95	await main()
96
97	if __name__ == "__main__":
98	asyncio.run(run())
99	```
100
101	### Streaming Responses
102
103	To stream agent output in real-time (e.g., for fluid UI or console applications), simply iterate over the `ChatResponse` object using an `async for` loop. The stream wrapper natively yields conversational `str` text tokens as they arrive, with zero network overhead:
104
105	```python
106	import asyncio
107	import sys
108	from google.antigravity import Agent, LocalAgentConfig
109
110	async def main():
111	config = LocalAgentConfig()
112	async with Agent(config) as agent:
113	# Returns instantly — does not block
114	response = await agent.chat("Write a short poem about space.")
115
116	async for token in response:
117	sys.stdout.write(token)
118	sys.stdout.flush()
119	print()
120
121	asyncio.run(main())
122	```
123
124	### Sugared Thoughts & Tool Call Streams (Advanced)
125
126	For more complex use cases, you can also stream internal model reasoning/thinking or intercept tool call dispatches in real-time using dedicated async stream properties:
127
128	```python
129	# 1. Stream reasoning/thinking deltas
130	async for thought in response.thoughts:
131	show_thinking_bubble(thought)
132
133	# 2. Stream strongly-typed ToolCall events
134	async for call in response.tool_calls:
135	s

People who liked this also viewed

6 picks

Type

antigravity-sdk-python

263kinstalls/mo · pypi2.6kstars961forks

View on GitHub

How to install antigravity-sdk-python?

google-antigravity/antigravity-sdk-python

$git clone https://github.com/google-antigravity/antigravity-sdk-python

Installs into the current project.

Summary

Files · 1

View on GitHub

README.md

1	# Google Antigravity SDK
2
3	The Google Antigravity SDK is a Python SDK for building AI agents powered by
4	Antigravity and Gemini. It provides a secure, scalable, and stateful
5	infrastructure layer that abstracts the agentic loop, letting you focus on what
6	your agent does rather than how it runs.
7
8	## Installation
9
10	```sh
11	pip install google-antigravity
12	```
13
14	> [!IMPORTANT]
15	> The Google Antigravity SDK relies on a compiled runtime binary that is
16	> included in the platform-specific wheels published to
17	> [PyPI](https://pypi.org/project/google-antigravity/). **Cloning this
18	> repository alone is not sufficient to run the SDK.** Always install from
19	> PyPI with `pip install google-antigravity` to obtain the binary.
20
21	## Quickstart
22
23	Get started by running one of the [`examples/`](examples/), such as the
24	`hello_world` example with:
25
26	```sh
27	export GEMINI_API_KEY="your_api_key_here"
28	python ./examples/getting_started/hello_world.py
29	```
30	## Gemini Enterprise Agent Platform (formerly Vertex AI)
31
32	To use the SDK with Gemini Enterprise Agent Platform (formerly Vertex AI),
33	configure `LocalAgentConfig` with `vertex=True` and specify your GCP `project`
34	and `location`.
35
36	By default, the SDK uses Application Default Credentials (ADC) for
37	authentication.
38
39	```python
40	from google.antigravity import Agent, LocalAgentConfig
41
42	config = LocalAgentConfig(
43	vertex=True,
44	project="your-gcp-project",
45	location="us-central1",
46	)
47
48	async with Agent(config) as agent:
49	response = await agent.chat("Hello!")
50	print(await response.text())
51	```
52
53	Alternatively, you can leave these fields unset in `LocalAgentConfig` and
54	export the environment variables instead:
55
56	```sh
57	# Either GOOGLE_GENAI_USE_VERTEXAI or GOOGLE_GENAI_USE_ENTERPRISE enable Vertex.
58	export GOOGLE_GENAI_USE_VERTEXAI=True
59	export GOOGLE_CLOUD_PROJECT="your-gcp-project"
60	export GOOGLE_CLOUD_LOCATION="us-central1"
61	```
62
63	Explicit kwargs always take precedence over env vars.
64
65	Ensure you have authenticated locally before running the agent:
66
67	```sh
68	gcloud auth application-default login
69	```
70
71	## Concepts
72
73	### Simple Agent
74
75	The `Agent` class is the easiest way to get started. It manages the full
76	lifecycle — binary discovery, tool wiring, hook registration, and policy
77	defaults — behind a single async context manager.
78
79	The `system_instructions` parameter is optional.
80
81	```python
82	import asyncio
83	from google.antigravity import Agent, LocalAgentConfig
84
85	async def main():
86	config = LocalAgentConfig(
87	system_instructions="You are an expert assistant for codebase navigation.",
88	# api_key="your_api_key_here",
89	)
90	async with Agent(config) as agent:
91	response = await agent.chat("What files are in the current directory?")
92	print(await response.text())
93
94	async def run():
95	await main()
96
97	if __name__ == "__main__":
98	asyncio.run(run())
99	```
100
101	### Streaming Responses
102
103	To stream agent output in real-time (e.g., for fluid UI or console applications), simply iterate over the `ChatResponse` object using an `async for` loop. The stream wrapper natively yields conversational `str` text tokens as they arrive, with zero network overhead:
104
105	```python
106	import asyncio
107	import sys
108	from google.antigravity import Agent, LocalAgentConfig
109
110	async def main():
111	config = LocalAgentConfig()
112	async with Agent(config) as agent:
113	# Returns instantly — does not block
114	response = await agent.chat("Write a short poem about space.")
115
116	async for token in response:
117	sys.stdout.write(token)
118	sys.stdout.flush()
119	print()
120
121	asyncio.run(main())
122	```
123
124	### Sugared Thoughts & Tool Call Streams (Advanced)
125
126	For more complex use cases, you can also stream internal model reasoning/thinking or intercept tool call dispatches in real-time using dedicated async stream properties:
127
128	```python
129	# 1. Stream reasoning/thinking deltas
130	async for thought in response.thoughts:
131	show_thinking_bubble(thought)
132
133	# 2. Stream strongly-typed ToolCall events
134	async for call in response.tool_calls:
135	s

People who liked this also viewed

6 picks

Type