Summarize a YouTuber's latest video into a Google Doc daily
A Make scenario runs at 9 AM, finds a YouTuber's newest video, has Hugging Face summarize it, and appends the summary to a Google Doc.
Testers had ChatGPT Agent navigate the WHO site, extract the latest data and organise it by indicator, country and trend; it often fails on logins or dynamic pages.
Testers had ChatGPT Agent navigate the WHO site, extract the latest data and organise it by indicator, country and trend; it often fails on logins or dynamic pages.
In ChatGPT Agent mode, the testers gave a single prompt asking it to retrieve the most recent publicly available World Health Organisation data. The agent navigated the WHO site by itself, pulled out the relevant figures and arranged them into a multi-tab spreadsheet. The tabs were sorted by indicators, countries and trends. The source doesn't describe any human steps beyond the prompt.
The source says the agent produced a clean multi-tab spreadsheet. It gives no figures on accuracy or time. It also warns that agent outputs often need reformatting or correction.
Agents often fail on sites with logins, unsupported browser states or dynamic single-page apps, so check the output and be ready to fix or reformat it.
Go to [data source website, e.g. WHO data portal] and pull the latest figures on [indicator] for [list of countries]. Build a spreadsheet with tabs organized by indicator, country and trend over time, with source links and the date of each data point. Finish when it's complete, and list any pages you couldn't access because of logins or dynamic content.
A Make scenario runs at 9 AM, finds a YouTuber's newest video, has Hugging Face summarize it, and appends the summary to a Google Doc.
An n8n workflow scrapes Reddit posts and comments via Bright Data, extracts concerns and themes with Gemini, and writes reports to Notion.
A user left a Gemini agent running overnight to analyze 170k LLM logs with the llm CLI and a free model, and it finished the job despite limits.