Best Debate Prep Tools & AI Card Cutters in 2026

By Zaid Anwar, Founder & Developer, DebateCardAI · Published July 28, 2026 in Guides

All prices and features checked July 28, 2026, against each tool's own public pages. We build one of the tools below.

The short version

If you do Policy, get Verbatim. It is free and it is what everyone else formats in.

If you want cards other debaters already cut, start with our own Logos. It is free on any account, it stays focused on the recent seasons where the live evidence is, and Deep Search finds a card from the argument you want to make, not just the tagline somebody else happened to write. haku.cards and PrepSync index more years of the same public wiki; if you need what one team read in one old round, they dig deeper.

If you want a card cut from a source you just found, us first here too. Cutting is the thing we built hardest, and the rest of this guide gives you enough to check that claim. Contention.ai, Deb8er, and Clusion cut as well.

If you have no money: every one of our features has a free tier you can actually use, not a demo. Five cuts a day, Logos uncapped, Rhetorify and Bracket Math free outright, trial runs of Speechify and Block Builder, no card on file. Verbatim, Collage, and Deb8er cost nothing either.

One more thing before the categories: cutting is where we started, not where we stop. Search, speech writing, block prep, and tournament math run on the same account, and the second half of this guide is mostly about tools that do one of those things.

Most people run two or three of these together. Almost nobody runs one.

Six things that all get called "debate prep tools"

Half the bad advice you get is somebody comparing across these rows.

A database and a cutter solve opposite problems. The database wins when someone already cut what you need. The cutter is for when nobody did, or when the tag they wrote does not match the argument you are running. Most serious tools now do both, which is why the rows overlap.

What to check before you pay anyone

Three tests, ten minutes each. We hold ourselves to the same three.

Can you get to the original article in one click? A card you cannot trace back to a live page is evidence you cannot produce when someone calls for it, which is the most expensive mistake in the activity.

Can you fix the highlighting fast? Every AI cutter gets the emphasis wrong sometimes, because the model does not know which warrant you are going for. The question was never whether it misses. It is how many clicks it takes you to fix.

Does the tool tell you it can be wrong? Clusion's terms say you must independently verify sources, accuracy, and citations, which is the right thing to put in writing. Here is why it matters: a 2025 study in JMIR Mental Health found 19.9% of GPT-4o's citations were fabricated outright, and of the ones pointing at real papers, 45.4% had bibliographic errors. Ours can be wrong the same way; that is exactly why every card we cut keeps the original source one click away, and why you read the source before you read the card in a round.

Run all three on a free tier before you give anyone money, us included.

Formatting and storage

Verbatim is a free, open-source platform for paperless debate, built as templates for Microsoft Word and Excel, currently on v6.0.0 for PC and Mac.

One line on their site explains more about this whole category than any pricing page does: Verbatim, openCaselist and everything else here is maintained by Aaron Hardy. openCaselist is the corpus almost every card database below is built on. One person, one project, most of the category downstream of it.

Get it if you do Policy, or you want your docs to look like what judges expect. It is free, so there is no reason not to.

Collage is a free Chrome extension with 3,000 users and a 4.9 rating. The developer describes it as more than a cutter: a way to cite, format, store, categorize, and search through the massive amount of information you've gathered. No AI anywhere, and storage is indefinite.

Get it if you research in the browser and want to grab, cite, and file without leaving the page.

Cut-It.Cards is a desktop download rather than a web app, with an AutoCite API that gets all the needed information from any website in just one click and exports to Google Drive, PDF, or Verbatim. No pricing shown.

Get it if you want a dedicated cutting app on a laptop and one-click citations.

Card databases

Index size is the number everyone advertises, and we think it is the wrong number to optimize for. Topics repeat, so the same card gets recut and redisclosed year after year; a bigger multi-year index mostly means more near-identical versions of the card you already found, and wading through fifteen copies of the same Futter cite is slower, not more thorough. The recent seasons hold the live evidence. That is why Logos stays focused on them and re-ingests weekly, and why we put the work into the searcher instead: getting you the best version of a card fast, and understanding the argument you are making well enough to find it even when the tagline says it differently. Read the big numbers below with that in mind.

haku.cards is straight about what it holds: 3.5 million cards indexed from openCaselist from 2013-2026 across hspolicy, hsld, nfald, hspf, and ndtceda, filterable by year, event, school, team, side, tournament, and round.

Read that first number twice. It is a fact about openCaselist, not a private library, which is equally true of ours. Their tournament and round filters go deeper than ours, and if you need what a specific team read in a specific round, use haku.

For everything else we would use ours. Logos indexes the same wiki, about a million cards of it, free and uncapped on any account, and adds Deep Search, where you describe the argument in plain English instead of guessing the tagline, plus a Claude connector so you can pull cards mid-conversation. Bigger index theirs, more useful per search ours.

Get it if you need to find what a specific team read at a specific tournament. Full breakdown: DebateCardAI vs haku.cards.

PrepSync claims 5.5 million pre-cut cards with new sources every two weeks, the biggest archive anyone here advertises. Pricing is $15/month, $99/season, or $29/month or $250/season for teams.

The free tier is the number to know: three card cuts, lifetime, not three a day. That is one afternoon. Their judge and debater lookup are both free and unlimited, and their judge lookup has no equivalent on our side.

Same call as with haku: theirs is the bigger archive, ours is the one we would rather search. Logos costs nothing on any account and never runs out, where PrepSync limits search on the free plan, and Deep Search finds an argument even when it is tagged in wording you would never have guessed.

Get it if you want the biggest archive and you research judges. Full breakdown: DebateCardAI vs PrepSync.

AI cutters and platforms

Contention.ai calls itself the first AI-powered debate search engine and takes the opposite architectural bet from everyone else: we never use wikis or repositories to find cards.

That is the cleanest split in this guide. haku, PrepSync, and our own Logos all lean on openCaselist. Contention.ai refuses to. Neither is automatically better. Wiki cards come pre-cut and round-tested; live-web cards are ones your opponent has never seen. They describe the tool as free but publish no tier table.

Get it if you want cards that are not already sitting on the wiki for your opponent to find.

Clusion has grown past prep into a team platform: rankings, evidence, coaching, judging, and tournament planning in one product system, plus a coaching and judging marketplace with listed rates in roughly the $30 to $90 an hour range. Their cutter is paste-based, and their index is 240,000+ cards.

Get it if your team wants rankings, judge info, and coaching sitting next to prep. Full breakdown: DebateCardAI vs Clusion.

Deb8er is free for all students, forever and built by a high-school student. It extracts up to 3 consecutive paragraphs per card and returns Word files in 30 to 60 seconds. Their reason for existing is worth reading: rich schools have coaches and libraries of debate cards, public-school debaters often have nothing.

Get it if you have zero budget and need real cards in Word. Check that three-paragraph cap against how you actually cut.

DebateCardAI

Ours, so here is the whole thing rather than the flattering third, plus the part where we tell you what we actually think.

What we think: if the job is turning a source into a card cut to your argument, we do it better than anything else in this guide. That is the thing we have built hardest and the thing we tune constantly, and it is the claim we would most like you to go test against whatever you use now. Ten minutes and one article settles it either way.

Cutting runs three ways. Single Card Snipe takes a URL and a claim and gives you one card. Batch Input runs in Automatic mode, where you approve each card as it lands, or Agent mode, where you load a stack, start it, and review at the end. Same model and prompts across all three, with tech or lay style and highlight density as settings, so a card comes out shaped for the panel you are reading to rather than generically highlighted.

Around that: Evidence Finder takes a claim with no source behind it, expands it into several searches across the web and academic databases, and ranks results by how cuttable they are, one claim or ten at a time. Logos searches over a million disclosed cards by tagline, by block header, or through Deep Search. A connector puts that library inside a Claude conversation. Rhetorify strips cards to what you say out loud, in-browser, no AI, works offline. Speechify writes a full constructive or rebuttal from a stack of cards. Block Builder walks from a strategy chat to a finished A2 block. Bracket Math turns a pairings link into your record and likely opponents. Debater Search looks up a competitor's Tabroom record and OpenCaselist page. The library files what you save by topic, and there is a Chrome extension.

The other thing we would point at is the pipeline, because no one else in this guide has it end to end: cut the card, condense it to what you actually read aloud with Rhetorify, then have Speechify write the speech from the stack. Most tools stop at the card. We think the half that comes after the card is where the time actually goes.

Free is 5 cuts a day, 5 Evidence Finder searches a month, three Speechify runs and two blocks to try them, and uncapped Logos, Rhetorify, Bracket Math, and library. No credit card. Pro is $8.99/month, the lowest monthly price in this guide, and unlimited for real prep; fair-use ceilings exist to stop automated abuse and no ordinary debater has come near one.

Two honest notes. Logos indexes the public caselists like most of this category, so treat our card count the way you should treat everyone else's. And there are real holes in our list: no flowing, no kritik shells, no practice rounds, no judge prep, no rankings, no coaching. Clusion does rankings and coaching, and both PrepSync and haku do judge research we do not.

Get it if you want research, cutting, and speech prep connected, and you cut more than three cards a week.

ChatGPT, Claude, Perplexity

Debaters use these constantly, mostly for the right things: pressure-testing an argument, finding the other side's best response, deciding whether a 40-page report is worth opening.

Evidence is the exception. See the JMIR numbers above. A chatbot will hand you a citation that looks perfect and is not real, and a card you cannot produce is treated as a card that never existed.

Use them for thinking. Not citations.

Which one to pick

Policy: Verbatim, with a cutter on top.

No money: Verbatim, Collage, and Deb8er are free, Contention.ai describes itself as free, and ours gives you 5 cuts a day. You can run a whole season on that stack.

Cheapest paid: ours, at $8.99/month.

Cards someone already cut: haku.cards for a specific team's round, PrepSync for the biggest archive. For everything else, our Logos, because it is free, uncapped, and searchable by argument.

Cards nobody has seen: any live-web cutter. Contention.ai or us.

Team logistics: Clusion.

Actually cutting the card: us. That is the one we will argue about.

Try before you pay

Take one article and one claim, run it through three cutters, and look at what comes back. Is the citation complete? Can you fix the highlighting? Does it link to the original? Ten minutes of that beats any comparison post, including ours, and it is how we would want you to decide.

If something here has changed, their site is right and we are wrong. Tell us and we will fix it.