Browserbeam Python SDK
Official Python SDK for the Browserbeam API — browser automation built for AI agents.
Installation
pip install browserbeam
Quick Start
from browserbeam import Browserbeam
client = Browserbeam(api_key="sk_live_...")
session = client.sessions.create(url="https://example.com")
# Page state is available immediately
print(session.page.title)
print(session.page.interactive_elements)
# Interact with the page
session.click(ref="e1")
# Extract with CSS, AI, and JS selectors combined
result = session.extract(
products=[{
"_parent": ".product-card",
"_limit": 3,
"name": "h2 >> text", # CSS selector
"price": ".price >> text", # CSS selector
"url": "a >> href", # CSS attribute
"rating": "ai >> the star rating out of 5", # AI selector
"in_stock": "js >> el.querySelector('.stock')?.textContent.includes('In stock')", # JS
}]
)
print(result.extraction)
# Close when done
session.close()
Async Support
from browserbeam import AsyncBrowserbeam
client = AsyncBrowserbeam(api_key="sk_live_...")
session = await client.sessions.create(url="https://example.com")
await session.click(ref="e1")
result = await session.extract(title="h1 >> text")
print(result.extraction)
await session.close()
Configuration
client = Browserbeam(
api_key="sk_live_...", # or set BROWSERBEAM_API_KEY env var
base_url="https://api.browserbeam.com", # default
timeout=120.0, # request timeout in seconds
)
Session Options
session = client.sessions.create(
url="https://example.com",
viewport={"width": 1280, "height": 720},
user_agent="Mozilla/5.0 ...", # omit for automatic rotation
locale="en-US",
timezone="America/New_York",
block_resources=["image", "font"],
auto_dismiss_blockers=True,
timeout=300,
)
Proxies
All sessions use a datacenter proxy by default (country auto-detected from the URL's TLD). No configuration needed. To customize:
# Use a residential proxy for a specific country
session = client.sessions.create(
url="https://example.com",
proxy={"kind": "residential", "country": "us"},
)
# Or bring your own proxy (overrides managed proxy)
session = client.sessions.create(
url="https://example.com",
proxy="http://user:pass@proxy:8080",
)
Available Methods
| Method | Description |
|---|---|
session.goto(url) |
Navigate to a URL |
session.observe() |
Get page state as markdown. Supports mode="full" for all sections. |
session.click(ref=) |
Click an element by ref, text, or label |
session.fill(value, ref=) |
Fill an input field |
session.type(value, label=) |
Type text character by character |
session.select(value, label=) |
Select a dropdown option |
session.check(label=) |
Toggle a checkbox |
session.scroll(to="bottom") |
Scroll the page |
session.scroll_collect() |
Scroll and collect all content |
session.screenshot() |
Take a screenshot |
session.extract(**schema) |
Extract structured data |
session.fill_form(fields, submit=) |
Fill and submit a form |
session.wait(ms=) |
Wait for time, selector, or text |
session.pdf() |
Generate a PDF |
session.execute_js(code) |
Run JavaScript |
session.close() |
Close the session |
Page Map & Full Mode
The first observe call automatically includes a page.map — a lightweight structural outline of the page's landmark regions (header, nav, main, aside, footer) with CSS selectors and descriptive hints. Use it to discover what content is available outside the main area.
res = session.observe()
for entry in res.page.map:
print(f"{entry.section}: {entry.hint}")
# nav: Home · Docs · Pricing
# main: Getting started with Browserbeam...
# aside: Related posts · Popular tags
To re-request the map on subsequent calls:
session.observe(include_page_map=True)
When you need content from all page sections (sidebars, footer links, nav items), use mode="full". The response markdown is organized by region headers:
full = session.observe(mode="full", max_text_length=20_000)
print(full.page.markdown.content)
# ## [nav]
# Home · Docs · Pricing
# ## [main]
# ...article content...
# ## [aside]
# Related posts · ...
Both parameters work identically with AsyncSession.
Session Management
Filter by status: "active", "closed", or "failed". Failed sessions ended with a fatal error; get returns error_code and error_message.
sessions = client.sessions.list(status="active")
failed = client.sessions.list(status="failed")
info = client.sessions.get("ses_abc123")
if info.status == "failed":
print(info.error_code, info.error_message)
client.sessions.destroy("ses_abc123")
Error Handling
from browserbeam import Browserbeam, RateLimitError, SessionNotFoundError
client = Browserbeam()
try:
session = client.sessions.create(url="https://example.com")
except RateLimitError as e:
print(f"Rate limited. Retry after {e.retry_after}s")
except SessionNotFoundError as e:
print(f"Session not found: {e.message}")
Documentation
Full API documentation at browserbeam.com/docs.
License
MIT
Release files for browserbeam 0.5.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| browserbeam-0.5.0.tar.gz | 11.5 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| browserbeam-0.5.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 25.5 kB
Release files / browserbeam-0.5.0.tar.gz
| Download URL | browserbeam-0.5.0.tar.gz |
|---|---|
| Size | 11.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
603ea0cbd50b85d8b334109fa606d2022eaa751eb5620478b7c1886688239ebc
|
|
BLAKE2b-256 checksum How to use checksums |
1eda40b769c883e19b19d7eb43efef7ac43befd591a6f46c4cc0cfa534fd9c01
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.3
|
Release files / browserbeam-0.5.0-py3-none-any.whl
| Download URL | browserbeam-0.5.0-py3-none-any.whl |
|---|---|
| Size | 13.9 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
82c72c5bf19e120883e9151fb94b5d5dde670f537b7e73b7e6788e6989438a39
|
|
BLAKE2b-256 checksum How to use checksums |
1052bcbd6abc2f0baf935068fe26138bf633e8f8e2dd36dac87e287bd6b4d257
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.3
|