Anchor Browser

Anchor is a platform for AI Agentic browser automation, which solves the challenge of automating workflows for web applications that lack APIs or have limited API coverage. It simplifies the creation, deployment, and management of browser-based automations, transforming complex web interactions into simple API endpoints.

This notebook provides a quick overview for getting started with Anchor Browser tools. For more information of Anchor Browser visit Anchorbrowser.io or the Anchor Browser Docs

Overview

Integration details

Anchor Browser package for LangChain is langchain-anchorbrowser, and the current latest version is .

Tool features

Tool Name	Package	Description	Parameters
AnchorContentTool	langchain-anchorbrowser	Extract text content from web pages	url, format
AnchorScreenshotTool	langchain-anchorbrowser	Take screenshots of web pages	url, width, height, image_quality, wait, scroll_all_content, capture_full_height, s3_target_address
AnchorWebTaskToolKit	langchain-anchorbrowser	Perform intelligent web tasks using AI (Simple & Advanced modes)	see below

The parameters allowed in langchain-anchorbrowser are only a subset of those listed in the Anchor Browser API reference respectively: Get Webpage Content, Screenshot Webpage, and Perform Web Task.

Info: Anchor currently implements SimpleAnchorWebTaskTool and AdvancedAnchorWebTaskTool tools for langchain with browser_use agent. For

AnchorWebTaskToolKit Tools

The difference between each tool in this toolkit is the pydantic configuration structure.

Tool Name	Package	Parameters
SimpleAnchorWebTaskTool	langchain-anchorbrowser	prompt, url
AdvancedAnchorWebTaskTool	langchain-anchorbrowser	prompt, url, output_schema

Setup

The integration lives in the langchain-anchorbrowser package.

%pip install --quiet -U langchain-anchorbrowser pydantic

Credentials

Use your Anchor Browser Credentials. Get them on Anchor Browser API Keys page as needed.

import getpass
import os

if not os.environ.get("ANCHORBROWSER_API_KEY"):
    os.environ["ANCHORBROWSER_API_KEY"] = getpass.getpass("ANCHORBROWSER API key:\n")

Instantiation

Instantiace easily Anchor Browser tools instances.

from langchain_anchorbrowser import (
    AnchorContentTool,
    AnchorScreenshotTool,
    AdvancedAnchorWebTaskTool,
)

anchor_content_tool = AnchorContentTool()
anchor_screenshot_tool = AnchorScreenshotTool()
anchor_advanced_web_task_tool = AdvancedAnchorWebTaskTool()

Invocation

Invoke directly with args

The full available argument list appear above in the tool features table.

# Get Markdown Content for https://www.anchorbrowser.io
anchor_content_tool.invoke(
    {"url": "https://www.anchorbrowser.io", "format": "markdown"}
)

# Get a Screenshot for https://docs.anchorbrowser.io
anchor_screenshot_tool.invoke(
    {"url": "https://docs.anchorbrowser.io", "width": 1280, "height": 720}
)

# Define a Pydantic model for the web task output schema
from pydantic import BaseModel
from typing import List


class NodeCpuUsage(BaseModel):
    node: str
    cluster: str
    cpu_avg_percentage: float


class OutputSchema(BaseModel):
    nodes_cpu_usage: List[NodeCpuUsage]


# Run a web task to collect data from a web page
anchor_advanced_web_task_tool.invoke(
    {
        "prompt": "Collect the node names and their CPU average %",
        "url": "https://play.grafana.org/a/grafana-k8s-app/navigation/nodes?from=now-1h&to=now&refresh=1m",
        "output_schema": OutputSchema.model_json_schema(),
    }
)

Invoke with ToolCall

We can also invoke the tool with a model-generated ToolCall, in which case a ToolMessage will be returned:

# This is usually generated by a model, but we'll create a tool call directly for demo purposes.
model_generated_tool_call = {
    "args": {"url": "https://www.anchorbrowser.io", "format": "markdown"},
    "id": "1",
    "name": anchor_content_tool.name,
    "type": "tool_call",
}
anchor_content_tool.invoke(model_generated_tool_call)

Chaining

We can use our tool in a chain by first binding it to a tool-calling model and then calling it:

Use within an agent

%pip install -qU langchain langchain-openai

from langchain.chat_models import init_chat_model

llm = init_chat_model(model="gpt-4o", model_provider="openai")

if not os.environ.get("OPENAI_API_KEY"):
    os.environ["OPENAI_API_KEY"] = getpass.getpass("OPENAI API key:\n")

from langchain_core.prompts import ChatPromptTemplate
from langchain_core.runnables import RunnableConfig, chain

prompt = ChatPromptTemplate(
    [
        ("system", "You are a helpful assistant."),
        ("human", "{user_input}"),
        ("placeholder", "{messages}"),
    ]
)

# specifying tool_choice will force the model to call this tool.
llm_with_tools = llm.bind_tools(
    [anchor_content_tool], tool_choice=anchor_content_tool.name
)

llm_chain = prompt | llm_with_tools


@chain
def tool_chain(user_input: str, config: RunnableConfig):
    input_ = {"user_input": user_input}
    ai_msg = llm_chain.invoke(input_, config=config)
    tool_msgs = anchor_content_tool.batch(ai_msg.tool_calls, config=config)
    return llm_chain.invoke({**input_, "messages": [ai_msg, *tool_msgs]}, config=config)


tool_chain.invoke(input())

API Reference:ChatPromptTemplate | RunnableConfig | chain

API reference

• PyPi
• Github
• Anchor Browser Docs
• Anchor Browser API Reference

Related

• Tool conceptual guide
• Tool how-to guides