takezJoin the beta

the take-settling app

Silly app.
Serious results.

Spin up hundreds of agents to settle any disagreement — with your business partner, your friends, or your family — and get a clear answer, backed by real proof.

how it works

Facts in. Answers out.

Two sides, one neutral panel, and a swarm of agents that only deal in what's true.

1

Each side gets a Corner

YouThem

Each side gets their own agent — a Corner that's on your side. You just talk to it in plain language, and it builds your points for you: your facts, your reasoning, even the hypotheticals.

2

Set the facts straight

Both sides review each other's personal facts — the things the AI couldn't possibly know on its own — and approve or deny each one.

3up to 500 agents

The Consensus steps in

Both locked cases go to one neutral panel nobody can talk to. It fans out a blind set of agents that research, fact-check, and loop until they agree.

4

Verdict

A result with a confidence score and the reasoning. No forced winner on a stubborn split — it leans and shows the dissent. Rebuttals need new evidence.

One partner points at the ad-spend chart while the other fumes

with your business partner

“We should double our ad spend before Q4.”

Make your case to your agent — your point of view, your facts, your evidence — then let hundreds of agents dig in and settle it.

Two friends debating a decision

with your friends

“Renting is just throwing money away.”

The debate that never ends. Drop in the numbers — rent, rates, how long you'll stay — and let hundreds of agents run the real math and call it.

A family deciding what's for dinner at the table

with your family

“Let's just get chicken nuggets for dinner.”

Plead your case — you're wiped, your stomach's been off all day. But a claim only counts if it holds up: if you never once mentioned feeling sick, it won't carry much weight.

every settle spins up a swarm

Pick how hard the agents dig.

Start free with 12 agents on every take. Crank it up when the answer really matters.

Quick

Free
12agents

Fast takes. Settled in a snap.

Deep

40agents

More corners, deeper digging.

Pro

120agents

Serious stakes, serious research.

Max

300agents

The full swarm. Nothing left unchecked.

BYOK — bring your own Anthropic key

Plug in your own key and pay just a small monthly fee. You cover your model usage; takez handles the agents and the orchestration.

Watch explanation video

the honest version

What’s the deal?

Labs like Anthropic and OpenAI have built genuinely remarkable language models. The real unlock for research isn’t one model answering once — it’s looping agents: models that draft, fact-check themselves, hunt for the gaps, and go again until the holes are filled.

takez runs on Anthropic’s models, orchestrated exactly that way. You bring the real facts and evidence; we spin up a swarm of agents to research, cross-examine, and pressure-test every claim.

Can you trust everything an LLM or the internet tells you? No. But give it solid facts and enough agents checking each other’s work, and the answer gets genuinely — usefully — close to the truth.