silverriver 0.1.37b9

SilverRiver SDK for advanced automation and AI-driven tasks

SilverRiver

SilverRiver is an SDK for advanced automation and AI-driven tasks.

Installation

You can install SilverRiver using pip:

pip install silverriver

Agentic API

Here's a basic example of how to use SilverRiver:

import os
import time

from silverriver import AgentChatInterface
from silverriver.client import Crux
from silverriver.interfaces import SupportedModes
from silverriver.interfaces.data_models import Observation, ChatMessage, HUMAN_ROLE


# Create a custom chat interface
class CustomChat(AgentChatInterface):
    def __init__(self, init_message):
        self.chat_history = init_message

    def send_message_to_user(self, message):
        # Implement this method to send messages to the user
        pass

    def wait_for_user_message(self):
        # Implement this method to wait for and return user messages
        pass


# Initialize the Crux client
client = Crux(api_key=os.environ["CRUX_API_KEY"])

# Create a chat instance with an initial message
chat = CustomChat(
    init_message=[ChatMessage(role=HUMAN_ROLE, message="On which website are you?", timestamp=time.time())]
)

# Create a browser session
browser_session, browser_obs, browser_meta = client.create_browser_session(
    start_url="https://www.example.com",
    chat=chat
)

# Create an observation
browser_obs = Observation(
    **dict(browser_obs),
    chat_messages=chat.chat_history,
    last_action_error="",
)

# Get an action from the AI agent
code = client.get_action(browser_obs, mode=SupportedModes.FLASH)

# Execute the action
browser_session.execute(code)

# Close the client when done
client.close()

Workflow Automation

SilverRiver now includes a command-line interface (CLI) for easier interaction. Here are the available commands:

Authentication

To authenticate with Crux and save the API token:

silverriver auth --api-key YOUR_API_KEY

This will save the configuration file in ~/.config/silverriver.

Record a Trace

To record a new trace:

silverriver record URL [-o OUTPUT]

• URL: The URL of the webpage to trace (must start with http:// or https://)
• -o OUTPUT, --output OUTPUT: The output filename for the trace (default: 'silverriver_trace')

License

This project is licensed under the MIT License.