01 — Cut
Read and cut in the same pane.
Paste a URL or attach a PDF. Read the article. Select and highlight, underline, or bold — Verba writes the cite, tag, and body in Verbatim-compatible formatting.
Verba
for debate
Verba
for debate
Verba is an evidence tool for competitive debate. Paste a URL or drop a PDF, select the sentences that matter, and ship a Verbatim-ready card in under a minute. An assistant trained on your cards helps you find the right warrant in every round.
Built for rounds
Cut cards in seconds. Search every warrant. Cite in one paste.
The assistant
Verba's assistant is trained on a corpus of over 100k competitive debate cards — plus every card in your library. Ask in natural language, get answers grounded in evidence you can actually read in round, with every claim traced back to a cite.
The assistant indexes a corpus of over 100,000 competitive debate cards plus every card in your library — tag, cite, body, and highlights — with a separate vector store per project so cross-contamination never happens.
/find, /compare, /answer, /block.
Every major open-evidence backfile, camp release, and topic starter. Indexed, embedded, and ready to recall the moment you ask.
Every card you save is embedded on save. No training step, no wait — it's searchable the moment you hit ⌘↵.
Your library indexes only your workspace. We never use your cards to train foundation models. Per-project sharing is explicit and off by default.
The evidence library
Filter by side, topic, or project. Type to search across tags, cites, and body text. Press / from anywhere in the app.
Pressure without an exit ramp produces the behavior it claims to deter
Mahbubani · Foreign Affairs · 3/24
Back-channel talks are live — a deal is possible now
Parsi · Quincy Institute · 2/24
Abstraction of sanctions masks the material violence of the dollar system
Ali · New Left Review · 1/24
Flexibility DA — curtailment collapses crisis response
Zarate · CSIS · 10/23
Taiwan Strait — de-escalation now avoids inadvertent war
Glaser · International Security · 1/24
Treating persons as means collapses the basis for any obligation
Korsgaard · Sources of Normativity · reprint
Threat construction produces the insecurity it purports to resolve
Campbell · Writing Security · reprint
Subsets — affirmative must defend all instances, not just one
[T] Subsets Masterfile · 2024
How it fits together
Paste a URL or attach a PDF. Read the article. Select and highlight, underline, or bold — Verba writes the cite, tag, and body in Verbatim-compatible formatting.
One searchable store across every project and tournament. Instant filter by side, topic, cite, date, or body word.
⇧E builds a round doc in priority order. Opens cleanly in Verbatim-style templates. No reformat.
Pricing
For individual debaters.
For teams sharing a backfile.
For coaches, camps, departments.
FAQ
Yes. Import .docx backfiles with highlighting, warrants, and shrinks preserved. Exports open in Verbatim-style templates without reformat.
No. It searches and ranks cards you've already cut. Warrants come from your evidence — the model is configured to refuse invented claims.
Your cards index only your workspace. We do not use them to train foundation models. Per-project sharing is explicit and you can turn it off any time.
Editing works offline. Library sync resumes on reconnect. The assistant falls back to on-device cached embeddings.
$9 per debater per month. First 30 days free. Cancel any time — your cards stay yours.