Honcho: Use Cases

Cross-topic synthesis

“How have central bank responses to inflation differed between the 2008 financial crisis and the post-COVID period? Draw on everything in my feeds.”

The assistant searches across years of indexed financial commentary, pulls relevant analysis from multiple sources, and synthesizes a comparison—threading together perspectives that would take hours to cross-reference manually. Every claim links back to a specific article, author, and date.

Track how expert opinion has shifted

“How has coverage of remote work evolved from the pandemic emergency through the return-to-office pushback? What shifted in the argument?”

The corpus preserves chronology. The assistant can trace a narrative arc—early emergency analysis, the productivity debate, the cultural backlash—and show how the same authors changed their positions over time. This is temporal reasoning that traditional search can’t do.

Rapid briefing assembly

“I have a board meeting tomorrow about AI regulation. Build me a briefing from my sources covering the EU approach, the U.S. approach, and the China approach.”

The assistant pulls articles across regions and regulatory frameworks, identifies the key analysts and their positions, and assembles a structured briefing with attributed quotes and contrasting viewpoints—in minutes instead of hours.

Pattern recognition across domains

“What patterns do my sources reveal about how different industries are adopting AI? Compare coverage of healthcare, finance, and education.”

The assistant searches across all indexed content by domain, identifies recurring themes—regulatory friction, workforce displacement, productivity claims—and surfaces cross-domain connections that no single-industry analyst would make.

Surface contrarian and minority views

“Find articles in my feeds where analysts disagreed with the consensus on interest rate direction. What arguments did they make, and how did things play out?”

The assistant searches for dissenting perspectives, maps them against subsequent events, and assesses which contrarian calls were right—a uniquely valuable capability when the corpus preserves both the prediction and the outcome.

Institutional memory

“Based on how the Financial Times has covered previous tech bubbles, what framework would their analysts likely apply to the current AI investment cycle?”

By synthesizing patterns across years of indexed analysis from a single source, the assistant can construct a plausible analytical framework for new situations—not prediction, but pattern-informed reasoning grounded in the source’s own editorial tradition.

Build a presentation from your research

“Design a presentation outline on the history of cryptocurrency regulation using only articles from my index.”

The assistant sequences your collected articles into a pedagogically sound structure—chronological narrative, key turning points, competing viewpoints—with each section grounded in specific sources you’ve already curated.

Network and influence mapping

“Which analysts appear most frequently in my feeds when the topic is semiconductor policy? What institutions do they represent?”

The assistant maps the expert network across your indexed content—identifying who covers what, which institutions are cited most, and how perspectives cluster. This is institutional knowledge that exists nowhere else in structured form.

Morning briefing

“Give me my morning briefing.”

The assistant calls get_digest to pull your personalized feed—entries from your favorited authors, sources, hosts, and saved searches—then summarizes them by topic.

Research a topic across your feeds

“What have I saved about tariffs?”

The assistant calls search_content and summarizes the results—pulling together perspectives from different sources you follow rather than doing a generic web search.

Compare your reading to the latest news

“What's the latest on the Supreme Court tariff case? How does it compare to what I've been reading?”

The assistant searches Honcho for your saved articles on the topic, does a web search for the latest developments, and gives you a combined briefing that highlights what's new vs. what you've already seen.

Save a web article with notes

“Save this article and tag it as 'research'—the key insight is that ocean acidification is accelerating faster than predicted.”

The assistant fetches the URL, extracts the content, and calls create_entry with your tags and a summary that includes your annotation.

Informed writing

“Draft a blog post about AI regulation using what I've collected.”

The assistant searches your Honcho index for everything tagged or related to AI regulation, combines it with current web research, and drafts something grounded in the sources you've already curated.

The search → favorite → digest loop

“Search for articles about AI safety.”

(reviews results)

“Good stuff. Add that as a favorite so I see new articles in my briefings.”

(later)

“Morning briefing please.”

The core workflow: search to find content you care about, favorite the sources and topics, then get a personalized briefing whenever you want.

Post analysis to the team

“Write up a summary of the Q4 earnings and post it to the research group.”

The assistant drafts the summary, then calls create_entry to publish it to the shared feed.

Share an article with commentary

“Share that Reuters article about the Fed with the research group, and add a note that it's relevant to our rate analysis.”

The assistant calls share_to_group to cross-post the entry. The original author is preserved, your note is prepended, and provenance metadata tracks who shared it and from where.

Catch up after time away

“I've been out for a week. Summarize what's happened in the research group.”

The assistant pulls recent group entries, reads through them, and gives you a synthesized briefing rather than just a list of titles.

Messaging

“Message the research group: the client meeting has been moved to Thursday.”

The assistant calls send_message to post to the group's shared feed. Messages, direct replies, and private channels all use the same infrastructure.

Discover and add sources

“Find RSS feeds for The Verge and add them.”

The assistant calls discover_feeds on the site URL, shows you the available feeds, and adds the ones you choose with add_source.

Build extraction rules conversationally

“Here’s the JSON from this API. Can you write extraction rules to pull out the articles?”

The assistant calls test_extraction_rules to preview how rules would extract entries from raw content (JSON, HTML, or XML) without saving anything. Iterate on the rules conversationally until they produce the right results, then save them.

AI memory

“Remember that the client prefers weekly reports on Mondays.”

The assistant calls save_memory to store the note. It can be recalled in any future conversation with recall_memory.

Send feedback

“Send a message to the feedback group: it would be great to have a dark mode option in the web UI.”

The assistant calls send_message to post your feedback. Suggestions, bug reports, and general comments are all welcome.

Deep Analysis