Use cases / Clawbrowser demo / 1:39
Research the Empire State Building on Wikipedia
Open Wikipedia in a managed browser, collect key Empire State Building facts, and return a source-linked summary.
Overview
A quick factual research task is more useful when the answer stays connected to its source. This demo asks an agent to inspect the Empire State Building article and summarize core facts from the visible Wikipedia page.
What happens in the video
The agent starts from a managed browser profile, opens the Wikipedia article for the Empire State Building, and returns a source-linked summary covering construction dates, height, location, and architecture. The article remains visible during the workflow, providing a direct source for the reported facts.
Before you start
- A managed browser profile available to the agent.
- A defined subject and a list of facts to collect.
Step-by-step guide
Practical steps for adapting the recorded task. The video summary above describes the demonstrated result; these checks guide your own run.
Open the source article
Start the selected browser profile, navigate to Wikipedia, and open the article for the Empire State Building rather than relying on an uncited model response.
Collect the requested facts
Read the article and extract the requested construction dates, height, location, and architectural context from the page available during the recorded session.
Return a linked summary
Present the facts in a compact answer and include the Wikipedia article URL so a reviewer can inspect the source and account for future page changes.
Try this prompt
Open the Wikipedia article for [subject]. Summarize [requested facts] from the page, distinguish dates and measurements clearly, and include the exact source URL so I can verify the result.
Adapt the placeholders to your task. This is an example prompt, not a transcript of the recording.
Results
- A concise summary of key Empire State Building facts.
- A direct link to the Wikipedia source used in the recording.
Questions about this workflow
Does the guide guarantee that the facts never change?
No. It documents the recorded workflow. Reopen the linked article when current facts or citations matter.
Can the same workflow use another source?
Yes. Replace Wikipedia and the subject in the prompt, then require direct source links in the output.
Put the workflow to work
Prepare the browser connection, then adapt the guide to your task and check its output.
Set up Clawbrowser →Related reading
- How to Build an AI Agent That Browses the Web (Python + CDP) →
See how browser access lets an agent inspect a live source before composing an answer.
- The Browser Runtime for AI Agents: Why Orchestration Isn't Enough →
Understand the runtime layer that gives an agent an observable, controllable browser session.
Explore the broader use case
Guide reviewed . Source: published video and description.
Related video guides
1:50Research featured HackerRank articles with Camoufox
Use a Camoufox profile to explore the HackerRank blog, collect three featured articles, and keep a source open for further reading.
Watch & read the guide
0:51Research top Reddit posts for developer keywords
Find relevant Reddit posts from the past year, collect engagement data, and export a community-research dataset.
Watch & read the guide
1:31Measure brand visibility in AI search
Compare brand mentions, placement, sentiment, competitors, and citations across ChatGPT, Perplexity, DeepSeek, and Microsoft Copilot.
Watch & read the guide