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.
Comet researched 50 HumanX speakers, surfaced themes and podcast guests, built a camera spec table, and cleaned up tabs.
Comet researched 50 HumanX speakers, surfaced themes and podcast guests, built a camera spec table, and cleaned up tabs.
Jason Howell used Perplexity's Comet browser, which can see all open tabs. For conference research, he gave it an email listing the first 50 HumanX speakers. Comet pulled the agenda, researched speaker backgrounds and surfaced themes. For cameras, he opened three product pages and asked for a markdown spec table. For tabs, he gave one prompt to group and clean up a messy browser.
Comet recommended the best speakers as podcast guests and built a clear camera table that ignored irrelevant tabs. It grouped articles and recipes, closed duplicates, and closed blank tabs after a follow-up. In a separate scheduling test it got the email recipient wrong.
Comet needs deep access to your tabs, calendar and accounts, which the author names as the trade-off. His verdict comes after only 24 hours of testing.
I'm attending [conference name]. Research the [number, e.g. 30] speakers listed on [speaker page URL]: for each, note their role, company, and topic, then summarize the main themes and suggest [5] who would make good podcast guests. Put it all in one table and then close any tabs you opened that are no longer needed, once I've seen the summary.
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.