NOVAUPTILET
Built by one team, tested on real work

Stop fixing the first draft

Most prompt collections are just screenshots of things that worked once. Ours are structured briefs: role, inputs, reasoning steps and output format. You spend less time coaxing the model and more time using the answer.

Browse the packsSee plans
Secure payments30-day money-back guaranteeFree revisions when models change
novauptilet.com/library/marketing-packOPEN
PromptsVariablesModel notesChangelog
01-positioning-brief.mdTESTED
02-launch-email-sequence.mdTESTED
03-ad-angle-generator.mdREVISED
04-landing-page-audit.mdTESTED
PROMPTS
48
MODELS
4
FORMATS
3
LAST FIX
6d ago
SAMPLE — ROLE BLOCK
You are a positioning strategist. Ask for: audience,
current pitch, top 2 competitors. Return a table of
3 angles with the proof each one needs.
12
Packs
1,925
Prompts
4
Models tested
30d
Refund window
The problem

Prompting looks easy until the output is wrong

Anyone can type a request into a chat box. The gap shows up later: the tone drifts, the structure changes between runs, half the answer is filler, and you end up rewriting it yourself. That is not a model problem, it is a brief problem.

Copy-pasted prompt lists rarely say what inputs they need or what they return. So you test, tweak, and test again — every single time you need the same thing.

✕Different output every run
✕No stated inputs or format
✕Breaks when the model updates
✕Written for one model only
✕Filler text you have to cut
✕No way to reuse it in a workflow
How it works

From purchase to usable output in minutes

1

Pick the pack for the job

Each pack page lists the prompts, the models it was tested on, and a full sample so you know what you are buying.

2

Download the files

Markdown, a Notion-ready doc and a JSON export land in your inbox straight after checkout. No dashboard to log into.

3

Fill the inputs and run

Every prompt states the variables it needs and the format it returns, so the second run looks like the first.

Written as briefs, not scraped from a thread

Tested before we ship

Every prompt runs against real briefs, code and campaigns. If it needs three retries to be useful, we cut it or rewrite it.

You get the files

Not a locked dashboard. Markdown, a Notion-ready doc and JSON exports you can drop into your own systems.

Revisions are free

Models drift. When behaviour changes we re-test, fix what broke and publish a new version. Owners get it at no extra cost.

Featured packs

1,925 prompts across 12 packs.

View all 12 packs

What we won't tell you

The prompt space is full of vague promises. Here is where we draw the line.

We don't claim AI will replace you.

Our prompts save time on repetitive framing. They don't make strategic decisions, and they won't read your mind.

One prompt does not fit every model.

ChatGPT, Claude, Gemini and Midjourney respond differently to the same wording. We test the variants instead of pretending otherwise.

Garbage in still means garbage out.

A good prompt can't fix vague input. Our prompts ask you the right questions so the model has something solid to work with.

Common questions

A folder of prompts, each written as a brief rather than a magic phrase. Every prompt tells the model what role to play, what inputs it needs, and what format to return. You also get a Notion doc, a markdown file, and a JSON export.

We test each pack on the models listed on its page. If a model needs different wording to behave, we include the variant. We don't pretend one prompt works everywhere.

When a model changes and a prompt breaks, we fix it and publish a new version. If you bought the pack, you get the new version free. No subscription required for packs.

Yes. If a pack doesn't do what the page says, email us within 30 days and we refund it. We might ask what went wrong, but we won't make you jump through hoops.

Read the full FAQ

Less prompting, more shipping

Buy once, own the files, and get every future revision as the models change.

Browse prompt packsCompare plans