https://github.com/lmnr-ai/index
A lightweight task engine for building stateful AI agents that prioritizes simplicity and flexibility.
https://github.com/lmnr-ai/index
ai ai-agents aiagent llm llm-agent workflow
Last synced: 23 days ago
JSON representation
A lightweight task engine for building stateful AI agents that prioritizes simplicity and flexibility.
- Host: GitHub
- URL: https://github.com/lmnr-ai/index
- Owner: lmnr-ai
- License: apache-2.0
- Created: 2024-11-30T19:34:02.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2025-04-07T23:16:42.000Z (29 days ago)
- Last Synced: 2025-04-08T00:24:34.372Z (29 days ago)
- Topics: ai, ai-agents, aiagent, llm, llm-agent, workflow
- Language: Python
- Homepage:
- Size: 1.49 MB
- Stars: 918
- Watchers: 10
- Forks: 36
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-github-repos - lmnr-ai/index - The SOTA Open-Source Browser Agent for autonomously performing complex tasks on the web (Python)
README
# Index
Index is a state-of-the-art open-source browser agent that autonomously executes complex tasks on the web.
- [x] It is powered by Claude 3.7 Sonnet with extented thinking. More models will be supported in the future.
- [x] Index is also available as a [hosted API.](https://docs.lmnr.ai/laminar-index/introduction)
- [x] You can also try out Index via [hosted UI](https://docs.lmnr.ai/laminar-index/introduction#hosted-ui) or fully [self-host the UI](https://x.com/skull8888888888/status/1910763169489764374).
- [x] Supports advanced [browser agent observability](https://docs.lmnr.ai/laminar-index/observability) powered by [Laminar](https://lmnr.ai).
prompt: go to ycombinator.com. summarize first 3 companies in the W25 batch and make new spreadsheet in google sheets.
https://github.com/user-attachments/assets/2b46ee20-81b6-4188-92fb-4d97fe0b3d6a
## Index API
Index API is available as [hosted api](https://docs.lmnr.ai/laminar-index/introduction) on the [Laminar](https://lmnr.ai) platform. Index API manages remote browser sessions and agent infrastructure. Index API is the best way to run AI browser automation in production. To get started, [sign up](https://lmnr.ai/sign-in) and create project API key.
### Install Laminar
```bash
pip install lmnr
```
### Use Index via API
```python
from lmnr import Laminar, AsyncLaminarClient
# you can also set LMNR_PROJECT_API_KEY environment variable
# Initialize tracing
Laminar.initialize(project_api_key="your_api_key")
# Initialize the client
client = AsyncLaminarClient(api_key="your_api_key")
async def main():
# Run a task
response = await client.agent.run(
prompt="Navigate to news.ycombinator.com, find a post about AI, and summarize it"
)
# Print the result
print(response.result)
if __name__ == "__main__":
asyncio.run(main())
```
When you call Index via API, you automatically get full browser agent observability on Laminar platform. Learn more about [Index browser observability](https://docs.lmnr.ai/laminar-index/introduction#tracing-with-laminar).
## Local Quick Start
### Install dependencies
```bash
pip install lmnr-index
# Install playwright
playwright install chromium
```
### Run the agent
```python
import asyncio
from index import Agent, AnthropicProvider
async def main():
# Initialize the LLM provider
llm = AnthropicProvider(
model="claude-3-7-sonnet-20250219",
enable_thinking=True,
thinking_token_budget=2048)
# Create an agent with the LLM
agent = Agent(llm=llm)
# Run the agent with a task
output = await agent.run(
prompt="Navigate to news.ycombinator.com, find a post about AI, and summarize it"
)
# Print the result
print(output.result)
if __name__ == "__main__":
asyncio.run(main())
```
### Stream the agent's output
```python
from index import Agent, AnthropicProvider
agent = Agent(llm=AnthropicProvider(model="claude-3-7-sonnet-20250219"))
# Stream the agent's output
async for chunk in agent.run_stream(
prompt="Navigate to news.ycombinator.com, find a post about AI, and summarize it"):
print(chunk)
```
### Enable browser agent observability
To trace Index agent's actions and record browser session you simply need to initialize Laminar tracing before running the agent.
```python
from lmnr import Laminar
Laminar.initialize(project_api_key="your_api_key")
```
Then you will get full observability on the agent's actions synced with the browser session in the Laminar platform.
### Run with remote CDP url
```python
import asyncio
from index import Agent, AnthropicProvider, BrowserConfig
async def main():
# Configure browser to connect to an existing Chrome DevTools Protocol endpoint
browser_config = BrowserConfig(
cdp_url=""
)
# Initialize the LLM provider
llm = AnthropicProvider(model="claude-3-7-sonnet-20250219", enable_thinking=True, thinking_token_budget=2048)
# Create an agent with the LLM and browser
agent = Agent(llm=llm, browser_config=browser_config)
# Run the agent with a task
output = await agent.run(
prompt="Navigate to news.ycombinator.com and find the top story"
)
# Print the result
print(output.result)
if __name__ == "__main__":
asyncio.run(main())
```
### Customize browser window size
```python
import asyncio
from index import Agent, AnthropicProvider, BrowserConfig
async def main():
# Configure browser with custom viewport size
browser_config = BrowserConfig(
viewport_size={"width": 1200, "height": 900}
)
# Initialize the LLM provider
llm = AnthropicProvider(model="claude-3-7-sonnet-20250219")
# Create an agent with the LLM and browser
agent = Agent(llm=llm, browser_config=browser_config)
# Run the agent with a task
output = await agent.run(
"Navigate to a responsive website and capture how it looks in full HD resolution"
)
# Print the result
print(output.result)
if __name__ == "__main__":
asyncio.run(main())
```
---
Made with ❤️ by the [Laminar team](https://lmnr.ai)