Regest AI

Ground AI workflows with source-verified quotes.

Regest is an MCP search tool that enables retrieval across thousands of hours of podcast and interview transcripts.

Try now →Workflows

Cheaper, Faster, More accurate analysis

Without Regest

Agents use web search to retrieve quotes
Less grounded agent responses and shallower analysis
More chances of misquotes
Slower and higher token usage for deep analysis

With Regest

Agents go direct to the source
More flexible querying improves analysis
Unearth interesting statements not already published on the internet
More reliable quotes
Source verified metadata for quoting
Hyperlinks and rendered audio

⭐ Testimonials

Our archive

Powered by thousands of hours of conversation

We have applied a novel indexing approach to thousands of the top audio content sources, making detailed search easy.

Podcasts

The Joe Rogan Experience
All-In with Chamath, Jason, Sacks & Friedberg
The Tim Ferriss Show
Pivot
Pod Save America
Honestly with Bari Weiss
The Journal.
Decoder with Nilay Patel
Lenny's Podcast
My First Million
On with Kara Swisher
Invest Like the Best
Plain English with Derek Thompson

Speakers

Elon Musk
Mark Zuckerberg
Jeff Bezos
Donald Trump
Barack Obama
Joe Biden
Vladimir Putin
Volodymyr Zelensky
Steve Jobs
Sam Harris
Reid Hoffman
Bernie Sanders

Broadcasters

CNBC
CNN
MSNBC
Bloomberg
Fox News
BBC
PBS NewsHour
Face the Nation
60 Minutes
Meet the Press
ABC News
HBO

and many more…

How it works

Our tools

Tool	Role
`find_ideas_on_topic`	Broad semantic search—explore what notable people have said about a topic when you do not have a specific argument in mind.
`find_evidence_for_argument`	Deep hybrid search for entailment—statements that support, complicate, or argue against a question or thesis.
`find_known_quote`	Fast, strict filters on episode and speaker—retrieve exact quotes when you already know the interview or speaker.
`find_quotes_about_person_or_event`	Keyword-grounded search—quotes about a named person, organisation, or event (profiles and retrospectives).
`find_quotes_by_speaker_type`	Clustered search by expertise—e.g. economist, tech CEO—with results grouped per speaker type.

The LLM agent decides which tools to use

mcp · regest

// e.g. one of five scenario tools
find_evidence_for_argument(
  "topic + context…",
  "your question or claim…"
)

Analyse, retrieve, and rank in one pass

Search steps complete

> preparing search inputs...> generating secondary analytical questions> hybrid retrieval across conversational index> entailment scoring and reranking> applying speaker and temporal signals> graph-based contextual boosts> diversity reranking> search complete

Conversational structure indexed as signal

Topic introductions and preceding questions point toward the statements that carry the argument.

Topic introduction

Preceding question

Retrieved statement

You need the Supreme Court to, you know, like the super friends, to come in in an emergency and rescue you from the villainous lower courts that are interfering with your rights. And so thats what it is. Has the emergency docket, has that been more packed than usual? Have we seen a lot more of them this year? And thats because there are many, many more challenges to Trump administration policies. And thats because there are many, many more provisional orders blocking those policies, which come from many, many more lawsuits. I mean, what does that meant for the court? How were they able to keep up with that new workload? Its a lot more work for them. And this is likely to continue over this summer. So it has changed their workload a lot. And justices have sometimes talked about that, sometimes viewing it as kind of annoyance, sometimes more substantively into saying, look, were making very consequential decisions based on very little information. And thats, you know, a problem. And maybe we should be having, you know, more hearings or find a way to make this at least more regular than it is. These cases are these big, impactful decisions, and they only have so much time to really review the information, and theyre not hearing oral arguments. Its just case after case after case. This year, the court has taken emergency cases over a variety of the Trump administrations actions. They include a ban on transgender service members in the military, the removal of officials from independent government agencies, and the repeal of temporary protected status for some migrants. In those cases, and others like them, the court has often ruled in favor of the administration, allowing the policy changes to take place.

Structured, source-verified output

search_result.json

{
  "statement": "It's a lot more work for them. And this is likely to continue...",
  "speakers": "ryan knutson and jess bravin",
  "podcast": "The Journal.",
  "episode_title": "The Supreme Court's Season Finale, Explained",
  "year": "2025",
  "source_audio": "https://pdst.fm/e/traffic.megaphone.fm/...",
  "themes": ["supreme court", "emergency docket", "judicial workload"],
  "keywords": ["Trump administration", "policy challenges", ...],
  "preceding_question": "How were they able to keep up with that new workload?",
  "full_context": { ... }
}