Skip to content

Open-source Python library for AI browser agents, plus Browser Use Cloud: stealth browsers and a hosted agent API.

revert extract changes

mertunsall1e3a36dbcee2
diff --git a/browser_use/tools/service.py b/browser_use/tools/service.py
index c53493501..47261d467 100644
--- a/browser_use/tools/service.py
+++ b/browser_use/tools/service.py
@@ -38,6 +38,7 @@ from browser_use.tools.views import (
ClickElementAction,
CloseTabAction,
DoneAction,
+ ExtractAction,
GetDropdownOptionsAction,
InputTextAction,
NavigateAction,
@@ -637,18 +638,21 @@ class Tools(Generic[Context]):
@self.registry.action(
"""LLM extracts structured data from page markdown. Use when: on right page, know what to extract, haven't called before on same page+query. Can't get interactive elements. Set extract_links=True for URLs. Use start_from_char if previous extraction was truncated to extract data further down the page.""",
+ param_model=ExtractAction,
)
async def extract(
- query: str,
+ params: ExtractAction,
browser_session: BrowserSession,
page_extraction_llm: BaseChatModel,
file_system: FileSystem,
- extract_links: bool = False,
- start_from_char: int = 0,
):
# Constants
MAX_CHAR_LIMIT = 30000
+ query = params['query'] if isinstance(params, dict) else params.query
+ extract_links = params['extract_links'] if isinstance(params, dict) else params.extract_links
+ start_from_char = params['start_from_char'] if isinstance(params, dict) else params.start_from_char
+
# Extract clean markdown using the unified method
try:
from browser_use.dom.markdown_extractor import extract_clean_markdown
diff --git a/browser_use/tools/views.py b/browser_use/tools/views.py
index d90f6cc33..48a4d2ff0 100644
--- a/browser_use/tools/views.py
+++ b/browser_use/tools/views.py
@@ -4,6 +4,16 @@ from pydantic import BaseModel, ConfigDict, Field
# Action Input Models
+class ExtractAction(BaseModel):
+ query: str
+ extract_links: bool = Field(
+ default=False, description='Set True to true if the query requires links, else false to safe tokens'
+ )
+ start_from_char: int = Field(
+ default=0, description='Use this for long markdowns to start from a specific character (not index in browser_state)'
+ )
+
+
class SearchAction(BaseModel):
query: str
engine: str = Field(
@@ -75,10 +85,6 @@ class UploadFileAction(BaseModel):
path: str
-class ExtractPageContentAction(BaseModel):
- value: str
-
-
class NoParamsAction(BaseModel):
model_config = ConfigDict(extra='ignore')