batonvm
Cloud browsers for agents, from Python. Every session is a real Chrome in its own Firecracker microVM, started in about 200 ms. A running session can move to another machine while your CDP connection stays open, be forked into copies that are logged in the same way, and be paused to storage and resumed later.
Quickstart
- Sign in at https://batonvm.vercel.app (email or GitHub) and create an API key.
- Install:
pip install batonvm playwright
export BATON_API_KEY=bk_...
Playwright only connects to the remote browser, so it needs no local browser download.
- Run:
from batonvm import Baton
from playwright.sync_api import sync_playwright
baton = Baton() # reads BATON_API_KEY
with baton.sessions.create() as session, sync_playwright() as p:
browser = p.chromium.connect_over_cdp(session.connect_url)
page = browser.contexts[0].pages[0]
page.goto("https://example.com")
print(page.title())
move = session.migrate() # the browser changes machines mid-script
print(f"moved {move.from_node} → {move.to_node}, paused {move.paused_ms} ms")
print(page.title()) # same page, same connection
The session ends when the with block does.
Async (browser-use, Playwright's async API)
import asyncio
from batonvm import AsyncBaton
from playwright.async_api import async_playwright
async def main():
async with AsyncBaton() as baton, async_playwright() as p:
async with await baton.sessions.create() as session:
browser = await p.chromium.connect_over_cdp(session.connect_url)
page = browser.contexts[0].pages[0]
await page.goto("https://example.com")
print(await page.title())
asyncio.run(main())
Move it, copy it, put it to sleep
copies = session.fork(3) # 3 new sessions: same pages, cookies, storage
for copy in copies.sessions:
print(copy.id, copy.connect_url)
session.pause() # stored; no machine holds it; up to 7 days
session.resume() # back on any machine; session.connect_url is new
session.save_profile("shop") # log in once...
later = baton.sessions.create(profile="shop") # ...start already logged in
API
Baton(api_key=None, base_url=None) |
client; api_key defaults to BATON_API_KEY (AsyncBaton for asyncio) |
baton.sessions.create(profile=None) |
start a browser → Session with connect_url |
baton.sessions.list() / get(id) / delete(id) |
|
session.migrate() |
move it to another machine → Migration(from_node, to_node, paused_ms, unreachable_ms) |
session.fork(n) |
copy it into n new sessions → Fork(sessions, paused_ms, errors) |
session.pause() / session.resume() |
store it and free its machine; bring it back, with a fresh connect_url |
session.save_profile(name) |
save its pages, cookies and storage; sessions.create(profile=name) starts from it |
baton.profiles.list() / delete(name) |
|
session.events() |
what happened to it: created, migrated, forked, paused, ended and why |
session.live_url() |
a short-lived URL for watching it |
session.close() |
end it (or use the session as a context manager) |
baton.usage() |
running sessions and minutes used today, with your limits |
Errors are BatonError with .status, raised as AuthError (401), NotFoundError (404), LimitError (429) or CapacityError (503). When every machine is full and a new one is starting, sessions.create() and resume() wait for it (about a minute) instead of failing.
Limits (free)
2 sessions running at a time · 15 minutes per session · 60 browser-minutes per day · 10 new sessions per minute · 5 paused sessions (kept 7 days) · 3 profiles.
Source: https://github.com/vedantxn/chromebox (MIT). TypeScript: npm install batonvm.
Metadata
Release files for batonvm 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| batonvm-0.2.0.tar.gz | 12.2 kB | Without provenance | Details for batonvm-0.2.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| batonvm-0.2.0-py3-none-any.whl | Python 3 | none | any | Details for batonvm-0.2.0-py3-none-any.whl |
Total release size: 22.7 kB
Release files / batonvm-0.2.0.tar.gz
| Download URL | batonvm-0.2.0.tar.gz |
|---|---|
| Size | 12.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
388826889344abececd4fa07f678c81e8f71a72a92e75930246d80628332eaaf
|
|
BLAKE2b-256 checksum How to use checksums |
6be56038dceda3aec3b1e979cae82ddce39de42ad1c0b9de0de1c8e007f032ce
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.14.6
|
Release files / batonvm-0.2.0-py3-none-any.whl
| Download URL | batonvm-0.2.0-py3-none-any.whl |
|---|---|
| Size | 10.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
a031b14c8380d2c3d82e41a4a4cc588f66d8d78a60020a7514cca193fe7ec2d2
|
|
BLAKE2b-256 checksum How to use checksums |
2b7ffc7af0bbe65e04a3143031eb7f1812b91226422334640c739230a92735cf
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/7.0.0 CPython/3.14.6
|
Security for batonvm 0.2.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.