Anthropic has brought its Claude Cowork system into the side panel of its Chrome extension, allowing users to run AI-powered research and document generation directly in the browser without additional setup.
The integration means skills, plugins, and connectors can now extract information from web pages and convert it into Excel files, PowerPoint presentations, or reports, though Anthropic warns users to avoid using it with sensitive financial or health information due to security vulnerabilities.
What happened
Anthropic integrated its Claude Cowork system (which creates files, slides, and reports from research) directly into the side panel of its Chrome extension. Skills, plugins, and connectors now work in the browser for the first time without extra setup, letting Claude handle research while Cowork produces Excel files, PowerPoint slides, or reports.
Why it matters
The integration eliminates friction for users who want to gather information from web pages and turn it into structured outputs — analytics dashboards become spreadsheets, Google Drive documents get organized, sales calls log automatically to Salesforce. Claude asks for confirmation before making purchases, and sessions are tied to your account rather than your device.
What to watch
The update is available now for all paid plans through the Chrome Web Store. Anthropic cautions that the extension remains vulnerable to prompt injection attacks and advises against using it with banking or health data.
Anthropic has expanded its Chrome extension by bringing Claude Cowork directly into the browser's side panel. Previously, the extension could read, click, type, and fill out forms on the pages you were viewing, but the workflow was limited to gathering information. Now, a full Claude Cowork session runs within that same panel, meaning users can research online and immediately have Claude generate Excel files, PowerPoint slides, or reports without leaving the browser or switching tabs.
The integration unlocks several practical workflows. Users can pull metrics from analytics dashboards, organize files stored in Google Drive, or automatically log sales calls into Salesforce—all without manual setup or copying data between tools. When Claude needs to take an action that might have financial consequences (such as making a purchase), it asks for user confirmation before proceeding. Sessions are tied to the user's account rather than a specific device, so the work persists across machines.
Anthropic has emphasized that skills, plugins, and connectors now work in the browser for the first time with no extra configuration required. This represents a significant reduction in friction compared to operating Cowork as a separate application.
The company has flagged a security limitation: the extension can be targeted by prompt injections—hidden instructions embedded in a webpage's content that could attempt to manipulate Claude's behavior. Like all LLM interfaces that interact with untrusted content, it remains a potential attack vector. Anthropic advises users not to use the extension with banking or health information. The update is available now for all users on paid plans and can be installed from the Chrome Web Store.
Anthropic's Chrome extension has long offered the ability to browse, click, type, and fill out forms on any webpage you're logged into—essentially automating low-level web interactions. Until now, that capability was limited to information gathering. The integration of Claude Cowork into the extension's side panel extends the extension's utility by allowing users to not only research but also synthesize that research into structured business outputs in a single workflow. This removes friction: users no longer need to switch contexts between a web browser and a separate Cowork interface to move from data collection to document creation.
The security posture reflects a broader tension in LLM-powered browser tools. Because the extension reads and interacts with the page you're viewing, it is theoretically exposed to malicious page content that could inject hidden instructions into Claude's reasoning (a "prompt injection" attack). Anthropic's explicit warning against use with banking or health data indicates the company is prioritizing user choice over locking down the feature, trusting users to assess their own risk tolerance based on the data they handle.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
Quanta Computer announced at its August 13 earnings call that it is raising its full-year 2026 capital expendi…
Salesforce spent $27 billion on stock buybacks in its first quarter ending in April—a record for the company—e…
Nebius Group and CoreWeave both reported Q2 2026 earnings this week showing strong AI compute demand
Bank of America analyst Vivek Arya lifted the 2030 server CPU total addressable market forecast to more than $…
Nebius, an AI cloud company listed on Nasdaq, has agreed to lease high-density NVIDIA-powered AI infrastructur…
Cerebras Systems fell over 18% in premarket trading after its second earnings report as a public company showe…
The AI news that matters, in one minute each morning.
Sign up free