With ChatGPT
In a custom GPT or Custom Instructions: paste the JSON contents into "About me". Done — every conversation will use it without you repeating yourself.
A JSON with everything we know about you, ready to drop into any LLM. So you stop explaining yourself every time you open a new AI.
The Portable Cognitive Profile (PCP) is a JSON containing your full profile: the six measured layers, the interpretation protocols and the cross-layer axes. It's not pre-cooked text — it's the raw data with instructions any AI can apply.
Paste the JSON as initial context. The AI reads it and starts treating you the way someone who'd known you a while would.
In a custom GPT or Custom Instructions: paste the JSON contents into "About me". Done — every conversation will use it without you repeating yourself.
In a Project: upload it as a knowledge file or paste it into the project instructions. Also works in the API in the system field.
If you have a custom agent (LangChain, OpenAI API, Anthropic API, local assistant), inject the JSON in the system prompt. Your agent will know who you are from turn one.
Two compatible products, not exclusive. When each one fits:
Standalone purchase — €249
You have your full profile in afini.ai and just want to export it once. One-off payment, downloaded JSON, ready to drop into any external AI. Doesn't include profile-aware chat inside afini.ai.
Professional plan — €49.99/mo
You want profile + ongoing chat inside afini.ai with Opus 4.7 + adaptive thinking + a quarterly export included. Every 90 days you get a fresh portable PCP, recalibrated with what you've added since.
If you have an active subscription, regenerate it from your dashboard as many times as you want. If you bought it standalone and have no subscription, we keep your last download for 12 months; after that you can re-buy at a reduced price.
The JSON is static: it reflects your profile at the moment of download. It doesn't expire. If you want recalibration on a schedule, the Professional plan includes a fresh export every 90 days.
The interpretation protocol is designed for Claude (Opus, Sonnet) and GPT-4 and later. On smaller or older models nuance is lost, but the raw information is still usable.
Only when you decide to use it. The JSON sits on your disk; when you paste it into an AI, that AI processes it under its own retention policies. If that worries you, local LLMs (Ollama, LM Studio) accept it too.
With an active subscription, yes — from your dashboard. Professional also delivers it automatically every quarter. A standalone purchase is a snapshot at the moment of download.
Because it generates ten hours of large-model compute for you: the narrative batch shipped with the JSON is computed on Opus, the protocols are versioned and signed, and the snapshot goes through scientific validation. Not a trivial export.