Whitepaper · Platform choice
The vendor lock-in you always wish you had
Every AI platform locks you into something. This paper shows where that lock-in should land, and how to test a vendor’s exit before you sign. Written for the people who choose.
What it covers
Lock-in is not optional. Where it lands is.
Any system that is useful collects your work: records joined across systems, rules for who sees what, tools people rely on. So you will be locked into something. Make it your own context, on your own servers, in files you can read.
- Layer 01
Model
Prompts, tuning, habits
- Layer 02
Data
Records and their links
- Layer 03
Permissions
Who may see what
- Layer 04
Logic
Schemas, tools, agents
- Layer 05
Runtime
Servers, keys, operations
Chapters
Six chapters, from the model to the exit
01 · Page 03All five layers
Something always builds up
Waiting and building it yourself are lock-in too.
02 · Page 06Model
The model is a setting
Why the model is the last thing to lock into.
03 · Page 08Data · Permissions
Your systems stay the system of record
Nothing migrated in, so nothing to migrate out.
04 · Page 11Logic · Runtime
What you build lives in your git
Schemas, endpoints and agents as files you hold.
05 · Page 13All five layers
The lock-in worth having
What stays sticky, said plainly.
06 · Page 15All five layers
Test the exit before you sign
Six questions for any vendor, and three steps.
Your data. Your infrastructure.