Skip to content

How it works Library Notes About

Enter app Follow the build

Loading page

A workspace for AI work you do more than once, built in the open.

In closed testing
Library free to use under CC BY 4.0

Product

How it works
Enter app
Newsletter

Library

Operators
Playbooks
Prompts
Knowledge

Read

Notes
Analysis
Build log
About

Elsewhere

X
LinkedIn
Encanta Digital

MyPromptVerse

© 2026 MyPromptVerse. Built by Encanta Digital.

NewsletterA few a month at most

Follow the build.

Notes from building MyPromptVerse in the open, sent when there is something worth reading. Written by Naeem, never more than a few a month.

Website

Email address

Occasional, practical, written by Naeem. Easy to leave. See privacy.

Ready for your email.

FromNaeem · MyPromptVerse

SubjectAI in October 2026: someone has to own the outcome

AnalysisAI in October 2026: someone has to own the outcome
Build logThe Quick button did nothing
GuidePrompt, Playbook, Operator, Knowledge: what goes where
New in the Library · PlaybookAPI Endpoint Reference

Occasional, practical, easy to leave.An example, from what is published now

01 In each issue

Three kinds of note, nothing else.

01piece so far
Analysis
Evidence-backed perspectives on AI developments and what they mean for the people building with it.
04entries so far
Build log
What we tried while building the app, what broke, and what we changed. Real traces, real numbers.
06guides so far
Guides
One practical idea per note on making AI work hold up: context, testing, review, when to automate.
19records, free to use
New in the Library
Operators, playbooks, prompts and knowledge as they are published from the Studio.

02 Recent notes

Read a few before you decide.

01AI in October 2026: someone has to own the outcomeWhat October's AI developments reveal about agent permissions, scientific verification and the responsibility that remains with the organisations deploying them.9 min read7 sectionsStrategy & Operations2 Nov 2026
02Prompt, Playbook, Operator, Knowledge: what goes whereThe four kinds of reusable AI setup, what belongs in each, and how they map to saved prompts, SKILL.md, Claude Projects and custom GPTs. Includes how to tell when a prompt you keep reusing should become a Playbook.7 min read7 sectionsStrategy & OperationsContent & Editorial13 Oct 2026
03The Quick button did nothingOur composer offered Quick, Structured and Deep. In a real run, all three changed only the wording of the prompt. This is how we found out and what replaced them.4 min read6 sectionsEngineering & DataProduct & UX13 Oct 2026
04We had 10,000 lines of multi-agent code. None of it called a model.Our app had a planner, scheduler, worker queue and tool registry for teams of agents. An audit found all of it ran as a local simulation. Here is what we kept and what we fixed first.4 min read5 sectionsEngineering & DataStrategy & Operations13 Oct 2026