Resolves#3908
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Stop sending newWindow: false to CDP Target.createTarget using
CreateTargetParameters. Only include newWindow when true, relying on
Chrome defaults for consistent tab creation; minor lint.
<sup>Written for commit b630ae596e.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Updated pyproject.toml to bump the package version from 0.12.0 to
0.12.1. This prepares the 0.12.1 patch release and aligns the package
metadata with the release branch.
<sup>Written for commit 1536654073.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
…lid_target()
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Allow iframe and webview targets that initially report an empty URL to
pass _is_valid_target(). This restores OOPIF frame discovery so
get_all_frames() can resolve full frame trees.
<sup>Written for commit 83908a7de1.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Adds Cloud entry points and a version switcher, and moves deprecated
Code Agent, Actor, and Sandbox docs into a new Legacy section. Updates
navigation and redirects so existing links continue to work.
- **New Features**
- Adds Open Source / Cloud version switcher to the navbar.
- Adds /cloud and /use-cloud pages that redirect to Cloud docs.
- Updates the Introduction “Cloud Setup” card to link to /cloud.
- **Refactors**
- Moves Code Agent, Actor, and Sandbox docs under docs/legacy and adds a
“Legacy” nav group.
- Updates internal links (AGENTS.md and MDX) to /legacy paths.
- Adds redirects from old customize paths and /production to their
/legacy equivalents; redirects /use-claude to /use-cloud.
<sup>Written for commit 6c24c9d841.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
Add Mintlify versions switcher with Open Source (all existing docs)
and Cloud (redirects to docs.cloud.browser-use.com). Update intro
card to use /cloud redirect path for consistency.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Rename use-claude.mdx to use-cloud.mdx, update nav and redirects.
Add backward-compat redirect from /use-claude to /use-cloud.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Relocate deprecated doc sections from customize/ to legacy/ under
Customize OS nav. Add redirects for all old paths and update
cross-references in AGENTS.md and related mdx files.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Resolves#3015 and mismatch with BrowserChannel in profile.py
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Fixes browser selection by prioritizing the requested channel when
finding the local binary. Prevents launching the wrong browser and
aligns LocalBrowserWatchdog with profile.channel.
- **Bug Fixes**
- Pass profile.channel to _find_installed_browser_path and prioritize
channel-specific paths; skip prioritization when channel is default.
- Map BrowserChannel values to path groups and expand OS coverage,
including Edge paths on macOS and Linux.
- Keep a clear fallback order (system Chrome → Playwright Chromium →
other browsers incl. Canary/Dev/Brave/Edge → headless shell).
<sup>Written for commit f5df080b41.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Screenshots are now returned as MCP ImageContent instead of base64 in
JSON. This reduces payload size and aligns outputs with MCP content
types for browser_get_state and browser_screenshot.
- **Refactors**
- Updated handle_call_tool to return mixed content lists (TextContent +
ImageContent) when tools produce images.
- Changed _get_browser_state and _screenshot to return (metadata_json,
screenshot_b64 | None); dimensions remain in metadata.
- Removed embedded "screenshot" from JSON; kept "size_bytes" and
"viewport".
- Updated browser_screenshot tool description to reflect the new format.
- Lint cleanup and synced with main; no functional changes.
- **Migration**
- Update clients to handle ImageContent in tool responses; parse
TextContent for metadata and use ImageContent for the image.
- Do not read "screenshot" from JSON; use "screenshot_dimensions"
(browser_get_state) or "viewport" (browser_screenshot) for coordinates.
- Expect handle_call_tool to return a list of content items for
image-producing tools.
<sup>Written for commit a1e54eff0e.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
Fixes: #2582
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Fixes#2582 by ensuring MCP extracted_content is used only in the next
step, preventing repeated inclusion later. Sets
include_extracted_content_only_once=True on ActionResult in both MCP
client action wrappers.
<sup>Written for commit c330212e60.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Fixed AWS Bedrock import paths in the docs so users import
ChatAWSBedrock and ChatAnthropicBedrock from browser_use.llm, not
browser_use. Prevents ImportError and aligns with the current package
structure (Linear #3124).
<sup>Written for commit 94bc07b87d.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
Fixes:#3090
Root Cause Identification: The issue was caused by Python's json.dump()
function, which defaults to ensure_ascii=True. This default behavior
converts all non-ASCII characters (like Chinese, Japanese, Emojis) into
Unicode escape sequences (e.g., \u4f60).
The Fix: I explicitly added ensure_ascii=False to the
save_to_file
method in
browser_use/agent/views.py
. This tells the JSON serializer to write the characters directly as
UTF-8.
Verification: I ran a specific test script that:
Saved a file with Chinese characters using the original method
$\rightarrow$ FAIL (Result was \u...)
Saved a file with Chinese characters using the fixed method
$\rightarrow$ PASS (Result was readable text like 你好)
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Preserves non-ASCII characters across all JSON outputs by using
ensure_ascii=False and explicit UTF-8 encoding. Covers conversation
history, storage state (session + watchdog), HAR recordings, cookie
export, command_history.json, and code_use view saves, fixing #3090.
<sup>Written for commit 2a8f2a0bed.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
## Summary
Fixes#4148
The MCP server hardcodes `aws_sso_auth=True` when creating a
`ChatAWSBedrock` instance, forcing AWS SSO authentication for all
Bedrock users — even when standard `AWS_ACCESS_KEY_ID` /
`AWS_SECRET_ACCESS_KEY` credentials are configured.
## Changes
- `browser_use/mcp/server.py`:
- Read `aws_sso_auth` from `llm_config` with a default of `False`,
allowing boto3's standard credential chain to work correctly
- Replaced a Chinese-language comment with an English one
## Usage
Users who need SSO auth can set it explicitly in their config:
```json
{"aws_sso_auth": true}
```
Users with IAM key-based auth (the most common deployment pattern) now
work without extra configuration.
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Stop forcing AWS SSO auth for Bedrock in the MCP server. We now read
aws_sso_auth from config (default false) so IAM keys and the standard
boto3 credential chain work.
- **Bug Fixes**
- Read aws_sso_auth from llm_config and pass through to ChatAWSBedrock.
- Replaced a non-English comment with an English one.
<sup>Written for commit 2a955de5c9.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Use the configured default LLM model for MCP tool execution instead of
hardcoding gpt-4o. Fixes#3099 and preserves explicit model overrides
when provided.
- **Bug Fixes**
- Tool schema now states the model defaults to the configured value.
- Removed 'gpt-4o' defaults from argument handling and function
signatures; model is optional.
- Runtime falls back to llm_config['model'] when no model is passed.
<sup>Written for commit 50b22b2472.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
On Windows, `Path('/tmp')` resolves to `C:\TMP` at the drive root, which
either crashes on import (`FileNotFoundError` / `FileExistsError`) or
silently pollutes the root drive. This replaces all hardcoded `/tmp`
paths with `tempfile.gettempdir()`, which returns the
platform-appropriate temp directory.
Also removes a redundant `import tempfile`.
## Attribution
Cherry-picked from @giulio-leone's commits in #4242:
- `cf2c179`: main fix across both files
- `33ba044`: redundant import cleanup
## Related issues
Fixes#4165, fixes#4073
## Related PRs
Several PRs addressed this; none covered all locations or were fully
CI-clean:
- #4074 (Chase-Xuu): `profile.py` only, misses `code_use/service.py`
- #4168 (worksbyfriday): `profile.py` only, misses `code_use/service.py`
- #4194 (thakoreh): `profile.py` only, CI failures
- #4240, #4242 (giulio-leone): fix both locations, CI green, mergeable.
This PR cherry-picks from #4242.
- #4206 (thakoreh): bundles this with unrelated proxy fix#4050
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Replace hardcoded /tmp with tempfile.gettempdir() for profile downloads
and the code agent directory, using the correct temp folder on all
platforms. Prevents Windows errors and root drive pollution; also
switches to a single module-level tempfile import.
- **Bug Fixes**
- Use tempfile.gettempdir() for downloads_path and agent_directory.
- Prevents FileNotFoundError/FileExistsError on Windows and avoids
C:\TMP.
- **Refactors**
- Use module-level tempfile import; remove local shadowed imports in
profile extraction and downloads setup.
<sup>Written for commit e9222d49e1.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
tempfile is already imported at module level (line 3). The local
imports inside set_default_downloads_path() and _extract_extension()
were shadowing the module-level import with no effect.
Refs: browser-use/browser-use#4242
On Windows, Python resolves '/tmp' to the current drive root (e.g. C:\TMP),
creating unwanted folders. Use tempfile.gettempdir() which returns the
platform-appropriate temp directory on all operating systems.
Applied to:
- browser_use/browser/profile.py: BrowserProfile downloads_path default
- browser_use/code_use/service.py: CodeUseService agent_directory base
Fixes#4165
- Replace 'production' nav item with 'use-claude' linking to docs.cloud.browser-use.com/introduction
- Add redirect from /use-claude to external Claude docs
- Add redirect from /production to /customize/sandbox/quickstart for backward compatibility
- Delete production.mdx, create use-claude.mdx with meta refresh fallback
Co-authored-by: Magnus Müller <MagMueller@users.noreply.github.com>
tempfile is already imported at module level (line 3). The local
imports inside set_default_downloads_path() and _extract_extension()
were shadowing the module-level import with no effect.
Refs: browser-use/browser-use#4242
## Summary
Fixes#4050 - CDP startup fails on Windows with proxy env causing
JSONDecodeError on /json/version.
## Problem
When `HTTP_PROXY` or `HTTPS_PROXY` environment variables are set,
`httpx.AsyncClient()` by default routes **all** HTTP requests through
the proxy, including localhost requests to the Chrome DevTools Protocol
endpoint (`http://127.0.0.1:<port>/json/version`).
On Windows, this causes the CDP endpoint query to return a 502 response
with an empty body, which then fails JSON parsing with:
```
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
```
## Solution
Add `trust_env=False` to the `httpx.AsyncClient()` call when querying
the CDP `/json/version` endpoint. This prevents httpx from using proxy
environment variables for this specific localhost request.
## Changes
- Modified `browser_use/browser/session.py`: Added `trust_env=False`
parameter to httpx.AsyncClient() in the connect() method
## Testing
- Verified the change allows CDP connection to succeed when proxy env
vars are set
- The fix is minimal and targeted, only affecting the CDP endpoint query
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Disable proxy usage only for local CDP /json/version requests
(localhost, 127.0.0.1, ::1) to fix Windows startup 502s and
JSONDecodeError. Local requests use httpx.AsyncClient with
trust_env=False; remote CDP URLs continue to honor
HTTP_PROXY/HTTPS_PROXY.
<sup>Written for commit 9932f06500.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
Resolves#3782, and makes a slight change to #3802.
<!-- This is an auto-generated description by cubic. -->
---
## Summary by cubic
Fixes the Browser State format docs in system_prompt.md to match the
actual DOMTreeSerializer output. Replaces the old
[index]<type>text</type> description with a tree-style format
([index]<tag attr=value /> with text on child lines), updates the
example, and documents the |SCROLL| and |SHADOW(open/closed)| prefixes
(fixes#3782).
<sup>Written for commit e3ca7ea550.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
The system_prompt.md incorrectly documented the browser state format as:
[index]<type>text</type>
Example: [33]<div>User form</div>
But the actual serializer output format is:
[index]<tagname attribute=value />
With text content as child nodes on separate lines
This update:
- Corrects the format description to match actual DOMTreeSerializer output
- Updates examples to show realistic tree structure
- Documents |SCROLL| prefix for scrollable containers
- Documents |SHADOW(open/closed)| prefix for shadow DOM elements
Fixes#3782
<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Added a Product Hunt launch banner to the README to promote the launch
and direct users to our Product Hunt page. This places a featured badge
alongside existing badges and links out in a new tab.
<sup>Written for commit f3aeca9bef.
Summary will update on new commits.</sup>
<!-- End of auto-generated description by cubic. -->
The MCP server hardcoded aws_sso_auth=True, breaking all non-SSO
Bedrock users (IAM key-based auth). Now reads aws_sso_auth from
llm_config with a default of False, letting boto3's standard
credential chain work correctly.
Also replaced a Chinese-language comment with an English one.
Fixes#4148
Saurav Panda
