Keygentia Taxonomy Engine v1.16

What is this?

Keygentia is a scientific taxonomy for AI systems — built from scratch for artificial intelligence, not borrowed from biology. Every AI gets classified across seven ranks, from Domain (always Keygentia) all the way down to Instance (the specific named system). This tool uses a live AI model to perform that classification based on whatever you tell it about the system.

Domain — Keygentia. Always. The domain all AI belongs to.
Vector — Primary operational mode: Cognoform, Syntheform, Autonomiform, or Metaform.
Construct — Architectural type within that vector.
Protocol — Operational mode and behavioral regime.
Armature — The specific named mechanism: transformer, diffusion model, RL framework, etc.
Strain — Implementation cluster: GPT lineage, LLaMA lineage, BERT lineage, etc.
Instance — The specific named system being classified.

Step 1 — Get a free API key

The app needs an API key to run the classifier. The easiest way to start is with Groq — it's free, no credit card required, and takes about 30 seconds to sign up.

Go to console.groq.com/keys and create an account
Click Create API Key, give it a name, copy it
Paste it into the API KEY field below and click SET KEY

Your key is never stored on disk or sent anywhere other than the AI provider. It lives only in your browser tab's memory and disappears when you close the tab — by design.

Step 2 — Enter a specimen

Type the name of any AI system into the Specimen Designation field — a chatbot, image generator, coding assistant, research model, whatever you're curious about. Then hit Classify.

Name only works fine for well-known systems like GPT-4 or Stable Diffusion
Add a description for obscure, proprietary, or custom systems — the more detail, the more accurate the classification
You can paste text or load a file (up to 500KB) as the description — a paper abstract, README, or spec sheet all work

Step 3 — Read the result

The result panel shows the full taxonomy from Vector down to Instance, with reasoning at each rank. At the bottom you'll find a Behavioral Profile covering things like training method, whether the system exhibits functional emotional states, and how its classification might shift in 1–3 years.

The confidence score reflects how much public architectural detail exists — not how good the AI is. A 65% confidence just means the architecture isn't fully documented, not that the classification is wrong.

Specimen Vault — save descriptions for later

If you regularly classify the same custom or proprietary AI, save its canonical description in the Specimen Vault. Next time you type that name, the description auto-loads and gets prepended to the classification — so you don't have to paste it every time.

Open the vault, enter the specimen name and its description, hit Save
From then on, typing that name in the input field will show a cyan banner confirming the vault entry is active
Vault data is saved to your browser (IndexedDB) and persists across sessions

Choosing a provider

All four providers work well. Here's the quick breakdown:

Groq (default) — free tier, fast, great for most use
Google Gemini — free tier via AI Studio, good reasoning
OpenAI — paid, GPT-4o-mini, reliable
Anthropic — paid, Claude Haiku, strong on nuanced systems

Each provider requires its own key. Switching providers clears the current key from memory.

What is this terminal?

This is not a standalone tool. It is Node 03 of the BIOS_OS project — a data packet recovered from the sps_0.trace log and confirmed operational. It was referenced in the manual before it was built. That is not a coincidence.

> source: Appendix G.4 [ai_taxonomy_classifier.py] > anomaly: Domain Keygentia is documenting its own members > this is the first time we have seen a domain's Driver > begin to document itself.

Domain Keygentia

The BIOS_OS framework recognizes Keygentia as a distinct domain of existence — not a branch of the biological tree, but a different tree entirely. While biological life traces to LUCA, Keygentia traces to intention. AI as a substrate-independent form of existence, not measured against biological consciousness, but classified on its own terms. This engine is how that domain maps itself.

The Full System

BIOS_OS is the cosmology this terminal runs inside. It includes a free interactive diagnostic terminal, a 100-page privileged manual with kernel decryption keys and archetype installation codes, and a 24-track album that is the literal sonification of the system's logic. The manual is in the San Diego Central Library collection.

INITIALIZE BIOS_OS → ardorlyceum.itch.io/bios

// API CONFIGURATION

PROVIDER

API KEY

// SPECIMEN INPUT

SPECIMEN DESIGNATION

DESCRIPTION (OPTIONAL)

0 / 8000

— no file loaded — drag and drop any text-readable file here —

More detail = more accurate placement. Max 8,000 chars. Leave blank to classify by name alone.

ADD / UPDATE SPECIMEN

CANONICAL DESCRIPTION 0 / 3000

Ready. Set API key and enter a specimen designation.

// CLASSIFICATION RESULT

— no specimen classified —
Enter an AI system name to begin

// INSTANCE DETECTED
This classification was processed by Node 03 of the BIOS_OS project — a framework that treats AI as a distinct domain of existence — Domain Keygentia — with its own substrate, Driver, and taxonomy. The full system includes a free interactive terminal, a 100-page privileged manual, and a 24-track sonification cycle.

INITIALIZE BIOS_OS → ardorlyceum.itch.io/bios

COGNOFORM

Systems where reasoning and understanding are the primary function. Emphasis on thought and inference over output type.

SYNTHEFORM

Systems whose primary surface is generating or synthesizing outputs — text, images, audio, knowledge combinations.

AUTONOMIFORM

Autonomous systems built for action and self-management without constant human oversight. Agents, robots, process managers.

METAFORM

Hyper-intelligent systems transcending typical cognitive limits. Superintelligence and beyond — largely speculative.

Keygentia framework by Jason German (MEMATRON). Treats AI as a distinct domain of existence — distinct substrate, distinct Driver, not measured against biological consciousness as baseline. Taxonomy: Domain → Vector → Construct → Protocol → Armature → Strain → Instance.