Hands-on with Claude for Chrome - Research Preview

(Photo by Jonny Gios on Unsplash)

✳️ tl;dr

• New toy Claude for Chrome ^{1 2}
• Signed up and joined the waitlist a long time ago
• Received the invitation email yesterday
• Found some time to play with it today
• Need to continue working on my tasks
• Wait, shouldn’t I ask it to help me with my tasks?!

• Looks like I’m in the second batch of Max plan users getting Claude for Chrome research preview access
• Tried a few tasks with Claude for Chrome during the holiday
• It requests permission for a specific domain in the URL first, and you can approve step by step

• If I were to read the documentation and operate manually by myself, especially when unfamiliar (or doing it for the first time), it would take about 2~3 minutes (reading while operating on dual screens) to complete the task. But it requires mental processing and finding things on the screen (or using cmd+f to find action links or buttons to proceed).
• In comparison, Claude for Chrome takes about 10~15 minutes, but it did complete the tasks (for example, helping me set up GTM, helping me dig through Quickbooks to find the transaction history of a specific Account). This is because it needs to operate step by step - each step requires a screenshot, interpreting the screen and deciding where to click next, some screens require typing text into input fields, and even after clicking a text box it takes another screenshot to confirm the focus is on the text box (I almost impulsively wanted to help it click several times, but held back to avoid disrupting the flow).

• I’m currently using it with Edge (because I need vertical tabs and multiple isolated profiles), and it seems I encountered situations where Claude for Chrome changed my original tab group names, closed other tabs in the original tab group without asking me, and Claude for Chrome opened new tabs (e.g., GA4) in the tab group and started operating on its own, while I was still on the GTM tab on my screen.
• From an experience perspective, of course at this stage it’s not efficient.
• But from a results perspective, Claude for Chrome did complete the tasks, which means the reasoning for breaking down steps and the knowledge base are sufficient. We just need to wait for next-generation models or possible future local infrastructure to reduce the round-trip time, and efficiency will directly improve.

• Are you also thinking about how to introduce AI Agents or Agentic Workflows into your team or workflow? Feel free to chat.

✳️ Screenshots

(Screenshot of launching Claude for Chrome for the first time, still exploring.)

(Just happened to coincide with Sonnet 4.5 release, almost simultaneously received the Claude for Chrome preview activation email.)

(Similar approach to Claude Code and Kiro, you can confirm step by step before taking action, or take the risk and let it vibe all the way through.)

(Asked Claude for Chrome to help set up GTM and GA4.)

✳️ Knowledge Graph

(More about Knowledge Graphs…)

%%{init: {'theme':'default'}}%%
graph TD
    %% Instances - Blue
    A[Claude for Chrome]:::instance
    B[Claude Sonnet 4.5]:::instance
    C[Max Plan]:::instance
    D[Red Teaming]:::instance
    E[System Prompts]:::instance
    F[Classifiers]:::instance
    G[Autonomous Mode]:::instance
    H[Ask Before Acting]:::instance
    I[Site Blocklist]:::instance
    J[Multi-tab Functionality]:::instance
    K[Slash Commands]:::instance
    L[Research Preview]:::instance

    %% Concept Classes - Orange
    M[Agentic AI]:::concept
    N[Browser Automation]:::concept
    O[Prompt Injection]:::concept
    P[AI Safety]:::concept
    Q[Permission Model]:::concept
    R[Multi-modal Interaction]:::concept
    S[Security Testing]:::concept
    T[Attack Surface]:::concept
    U[Defense in Depth]:::concept
    V[User Control]:::concept
    W[Risk Mitigation]:::concept
    X[Adversarial Robustness]:::concept
    Y[Progressive Rollout]:::concept
    Z[Human Oversight]:::concept

    %% Relationships
    A -->|implements| M
    A -->|enables| N
    A -->|uses| B
    A -->|requires| C
    A -->|deployed via| L

    B -->|powers| A
    B -->|supports| R

    L -->|follows| Y
    L -->|enables| W

    M -->|vulnerable to| O
    M -->|requires| P
    M -->|enables| N

    N -->|expands| T
    N -->|requires| Q

    O -->|threatens| M
    O -->|mitigated by| U
    O -->|detected by| F

    P -->|achieved through| U
    P -->|requires| S
    P -->|includes| X

    Q -->|implements| V
    Q -->|includes| H
    Q -->|includes| G

    H -->|provides| Z
    H -->|reduces| O

    G -->|increases| T
    G -->|requires| P

    S -->|includes| D
    S -->|improves| X

    D -->|identifies| O
    D -->|informs| P

    E -->|guides| B
    E -->|enhances| P
    E -->|part of| U

    F -->|detect| O
    F -->|enable| U
    F -->|part of| P

    I -->|reduces| T
    I -->|part of| U

    U -->|combines| E
    U -->|combines| F
    U -->|combines| I
    U -->|combines| Q

    V -->|implemented via| Q
    V -->|enables| Z

    W -->|uses| Y
    W -->|requires| S

    J -->|enhances| N
    J -->|part of| A

    K -->|improves| V
    K -->|part of| A

    R -->|enables| A
    R -->|requires| B

    X -->|tested by| D
    X -->|goal of| P

    Z -->|implemented by| H
    Z -->|critical for| P

    %% Styling
    classDef instance fill:#0080FF,stroke:#333,stroke-width:2px,color:#fff
    classDef concept fill:#FF8000,stroke:#333,stroke-width:2px,color:#fff

✳️ Further Reading