Prompts

Prompt collections built for specific tasks and checked against real use, not demos. Delivered as files you own outright. Version them, copy them, and edit them however you like.

Licensing: One-time, per organization

Delivery: Plain-text files

Book a free call

Overview

A prompt library is a set of tested prompts for work you repeat: drafting client emails in your voice, turning meeting notes into your summary format, pulling fields out of documents, or checking a draft against your standards. Each prompt is written for a specific task and tested on real examples before delivery.

Prompts need upkeep

Models change, and a prompt that works well on one version can behave differently on the next. OpenAI's prompting guide for GPT-4.1 notes that the model follows instructions more literally than earlier versions, so existing prompts may need adjusting. Every prompt we deliver carries test cases, so you can recheck it quickly after a model update.

Delivered as files

Prompts arrive as plain-text files, organized by task and ready for version control or a shared folder. Use them in any AI tool, copy them, and edit them however you like.

What each prompt includes

Every prompt in a library ships with the same structure, so anyone on your team can use it, test it, and change it.

client-summary.md

Purpose: The task it does and when to use it.
Inputs: What to paste or attach, and in what form.
Instructions: The prompt itself, written for your standards and voice.
Examples: Good output from real cases, so the model and the reader know the target.
Test cases: Inputs with expected results, to recheck the prompt after a model update.
Model notes: Which models it was tested on and anything that differs between them.
Version: A change log, so you can see what changed and roll back.

What you receive

Plain-text prompt files, organized by task
Test cases for every prompt
Model notes and version history
A short guide to adapting prompts for new tasks

Questions

Which models do the prompts work with?

Each prompt lists the models it was tested on, and its test cases show how it holds up on others.

Can you write prompts for our specific work?

Yes. Custom libraries start from your examples, your standards, and your voice.

How do we keep them up to date?

Rerun the test cases when your model changes. If results drift, adjust the prompt or ask us to.

Do the prompts connect to any outside service?

No. They are text files. Nothing runs until you paste or load a prompt into your own AI tool.