Chat is free. The review layer is $19 a month.
A wrong contract, a rushed hire, a launch that flops — the price of getting it wrong dwarfs the price of pressure-testing it first. Chat is free, with no account and no card. When the call is one you can’t easily undo, the Multi-Agent Team takes it: rival frontier models debate it, a separate audit signs it off, and you see exactly what to verify. Plus includes 16 of those every month, for one person — there are no seats to buy.
Chat costs nothing and needs no account. Sign up free and both your first specialist consultation and your first full Multi-Agent run are on the house.
How much does Decidi cost?
Chat is free and needs no account — and you choose which lab answers it. Plus is $19 a month: unlimited chat on the Standard models, unlimited specialist minds, panels, and 16 full runs of the Multi-Agent Team every month — each one debated across rival frontier models and signed off by the proprietary Final QA audit. Pro is $49 and adds the Deep models in the chat itself, working teams, Deep depth and 44 full runs a month. A free account keeps the chat and includes one specialist consultation. Cancel at any time.
That is about $1.19 a run on Plus, and $1.11 on Pro. Set against a contract you can’t unwind or a hire that doesn’t work out, a year of either plan is a rounding error — a debated, audited, signed-off call is the cheapest insurance you’ll buy.
One model per lab on each rung, picked in the chat itself. The Quick rung is free for everyone, with no account at all. A plan doesn’t only buy runs — it changes the model that answers every message you send, and every reply is stamped with the model that wrote it.
- GPT-5.6 Luna
- Gemini 3.5 Flash Lite
- Claude Haiku 4.5
- Grok 4.1 Fast
- GPT-5.6 Terra
- Gemini 3.7 Flash
- Claude Sonnet 5
- Grok 4.3
- GPT-5.6 Sol
- Gemini 3.1 Pro
- Claude Opus 5
- Grok 4.5
Choose a plan
What each plan opens · cancel anytimeUnlimited chat and specialist minds, with the full Multi-Agent Team on tap.
Everything in Plus, plus teams that produce the deliverable and Deep depth.
Top up, only if you outrun your plan
A first top-up — a handful of extra runs on the meter.
Enough for a busy month that runs past your plan.
For sustained work — a long run of full debates and deliverables.
The most credits per dollar — for daily audits and strategy work.
Run counts are honest approximations. A full run is one question taken from your first message to a signed-off verdict; the exact draw depends on how many minds sit at the table and how long the debate runs, and a Deep run (Pro) draws about 5Standard runs’ worth of the meter. Top-ups stack on top of your plan and never expire — they are there for the month that runs long, not the way to buy Decidi.
You buy runs, not credits. Underneath, every run is metered against the real, live cost of the models it uses, plus the orchestration and the audit we run on top — which is why full runs carry a real number instead of the word “unlimited”. Three of them a day costs more than a $19 plan collects, and we would rather publish a true figure than pad the price or throttle you quietly and call it fair use. Chat and specialist minds are uncapped because they genuinely are cheap. If a month runs long, a top-up from $5 adds to the same meter.
Keeping a seat at each lab — ChatGPT, Claude, Gemini, Grok — runs to about $20 a month each, before any of them will argue with the others. Plus opens a frontier model from all four in one chat for $19, and a full run makes them debate each other — which no number of separate subscriptions will ever do.
Your run count is a published number computed from the plan itself — it cannot quietly change with demand, and there are no compute points or platform fees to decode.
Cancel in one click from your account. You keep the plan to the end of the period you paid for, and everything in your wallet stays yours — nothing expires on the way out.
Every run is scored on delivery. A result that genuinely misses the brief is refunded automatically — scored by an independent model, not by a support queue.
Beyond the verdict, “Build it out” produces the finished file itself — a document, a dataset, an image, a short video, or a separate page for every row of a file. It is included in your plan and draws on the same monthly allowance as your full runs, so it helps to know what each one weighs:
A short video genuinely costs that much more to render than a page of text — that is why it is weighed higher. As with everything, the estimate is shown first and only what is actually produced draws on your plan.
Honest pricing, no traps
Take a plan for as long as it earns its place, and cancel in one click whenever you like. Credits you have already bought stay yours, and there are no surprise charges either way.
Buy once and use them whenever. Unused credits roll over forever — no monthly reset, no “use it or lose it”.
Your plan states exactly how many Multi-Agent runs it includes, chat and specialist consultations are uncapped, and nothing runs until you accept it. You’re only charged for work that actually ran.
Add your company name, address and VAT/Tax number and download a proper invoice from your billing history — issued automatically on every purchase.
Reach a human, not a black hole. If something went wrong with a purchase, email us and we’ll make it right.
Refunds are automatic where they can be: a run that fails part-way returns every credit it held, and a verdict that fails our own Final QA quality bar refunds itself — no email needed. The full wording is in the terms.
Pricing questions
How much does Decidi cost?
Chat is free and needs no account. Plus is $19 a month and includes unlimited chat on the Standard models, unlimited specialist minds, panels, and 16 full runs of the Multi-Agent Team every month. Pro is $49 and adds the Deep models in the chat itself, working teams, Deep depth and 44 full runs a month. A free account keeps your conversations, and includes one specialist consultation and your first full Multi-Agent run — both on the house. No card to start, and you can cancel a plan at any time.
Which AI models can I chat with?
Free chat runs on the Quick models — GPT-5.6 Luna, Gemini 3.5 Flash Lite, Claude Haiku 4.5, Grok 4.1 Fast — and you choose which lab answers, with no account. Plus opens the Standard models in the chat itself (GPT-5.6 Terra, Gemini 3.7 Flash, Claude Sonnet 5, Grok 4.3), and Pro opens the Deep ones (GPT-5.6 Sol, Gemini 3.1 Pro, Claude Opus 5, Grok 4.5). The picker shows all twelve, so you can always see what a plan opens, and every reply is stamped with the model that wrote it. A full Multi-Agent run sets rival models from different labs against each other whichever plan you are on.
What counts as one run?
One question taken from your first message to the signed-off verdict: the minds are chosen, they debate across rival frontier models, live research grounds the facts that move, and the Final QA audit signs the result off. Re-reading it later or talking the result over in chat costs nothing more — chat is free on every plan.
Why 16 and 44 runs rather than unlimited?
Because unlimited would not be true. A full run puts rival frontier models through several rounds and then an independent audit, and that costs real money every time — three of them a day costs more than a $19 plan collects. The honest options are to pad the price for a month almost nobody has, to throttle quietly and call it fair use, or to state a real number. We state the number. Chat and specialist minds are genuinely unlimited because, at the pace a person actually types, they cost little enough to absorb — even on the frontier models a plan opens in the chat. Unlimited there means any human pace: the only ceiling is an anti-abuse one set far above real use, there so scripted misuse can’t degrade the service — stated here for the same reason we state the run numbers.
Does building the finished deliverable cost extra?
No. After the team agrees what to make, “Build it out” produces the actual file, and it is included in your plan — it draws on the same monthly allowance as your runs. The weights differ honestly: a document or report is one to two runs, an image up to a third of one, and a short video fifteen to twenty, because video genuinely costs that much to render. The estimate is shown before it runs, and only what is actually produced draws on your plan.
What if I outrun my plan?
Top-up packs start at $5 and add to the same meter your plan already runs on. They never expire and they roll over, and members pay less for them — 10% off on Plus, 20% off on Pro. Most months you will not need one; they exist so a heavy month costs a few dollars rather than forcing an upgrade.
Do I have to subscribe?
Not to use Decidi — chat is free, and a free account adds a specialist consultation on the house. A plan does two things: it moves your everyday chat up the model ladder (Standard on Plus, Deep on Pro), and it opens the work where minds argue with one another — panels, the full Multi-Agent Team and, on Pro, working teams and Deep depth. That is where the real model cost sits, which is why it sits behind a plan. Cancel anytime.
Do I get an invoice?
Yes. Prices are in USD. A receipt is issued for every purchase, and you can download a proper VAT invoice — with your company name, address and tax number — from your billing history. If your card isn’t in USD, your bank applies its own conversion rate.
Can I try it before paying?
Yes — twice over. Chat is free with no account at all. And a free account includes one specialist consultation and your first full Multi-Agent run, both on the house — so you experience the exact product a plan buys before deciding whether it is worth it.
Chat is free, no card, no sign-up · create an account and a specialist consultation is on the house