AI reality, Onchain necessity and Token necessity

Every project card on DecAIHub contains an AI thesis fit section that answers three questions. This page explains what each of them means and how to read them.

In short: these are three verifiability axes. They show how well-supported a project's claims about AI, blockchain, and its token are — and they say nothing about price or investment attractiveness.

AI reality — what is genuinely AI here

Question: does the project have a real AI component that processes data and produces results (recognizes, recommends, predicts, generates), or has no verifiable evidence of a working AI component been found so far.

What is checked (progressively):

AI_0 — is there at least a Tier-1 declaration of AI intent (roadmap, official update).
AI_1 — is there a description of what exactly the AI does.
AI_2 — are there demos or working examples.
AI_3 — are there quality metrics and benchmarks.
AI_4 — is there a technical description, architecture, or model versions.

The resulting ai_score ranges from 0 to 3. If ai_score = 0, then DecAI Fit is automatically 0: without an AI component confirmed by Tier-1 sources, the project does not receive a DecAI Fit above 0 as of the verification date.

Onchain necessity — what is genuinely onchain

Question: which part of the project actually operates on a blockchain, and which part runs on a conventional server.

What is checked:

OC_1 — are there contracts, addresses, and a network that can be verified.
OC_2 — is on-chain usage more than just a token; does it include useful logic (access control, accounting, rules, DAO).
OC_3 — is on-chain critical: if the blockchain were replaced with a conventional database, would the product lose a key property (verifiability, censorship resistance, trustless settlement).

The resulting onchain_score ranges from 0 to 3. The higher it is, the stronger the project's dependence on blockchain.

Token necessity — why the token is needed

Question: what specific role does the token play in the product and how well-supported is that role.

What is checked:

TK_0 — no token exists (if so, the project receives token_score = 2: the absence of a token is not penalized).
TK_1 — the token is used to pay for services.
TK_2 — the token is needed for incentives (rewards, staking, slashing).
TK_3 — the token is needed for governance (voting).
TK_4 — the token is replaceable (could be substituted with USDC or a subscription without losing functionality).

The resulting token_score ranges from 0 to 3. The TK_4 flag (replaceability) lowers the score: if the token is easily replaceable, its necessity is in question.

How to read special cases

Not every card contains clear answers on each axis. Here is how to interpret the status values:

Value in the card	What it means
`unknown`	No information found in any Tier source. This does not mean the technology is absent — it means evidence has not been found yet.
`no token`	The project has no native token. This is a neutral fact, not a shortcoming.
`token use case not established`	A token exists, but its specific function within the protocol has not been confirmed by Tier-1 sources.
`positioning only`	AI is mentioned in marketing or positioning, but no Tier-1 evidence of a working AI component has been found.
`none`	Nothing has been detected for this aspect. Like `unknown`, this is an evidence-reading mode, not a statement about the absence of technology.

How the three axes combine into DecAI Fit

Scores across four components (AI reality, Onchain necessity, Token necessity, and Evidence) are summed and converted into the final DecAI Fit on a 0-to-6 scale:

0 — AI reality and AI intent not detected based on the sources found.
1–2 — low alignment with the Decentralized AI thesis.
3–4 — moderate alignment.
5–6 — high alignment.

Governance caps also apply: for example, without Tier-1 evidence for AI or onchain, the score cannot exceed 3.

Do not confuse with

This is not a financial assessment. AI reality, Onchain necessity, and Token necessity measure the verifiability of a project's claims, not its market value.
A low score is not a verdict. A project may be at an early stage where evidence has not yet emerged. The card records the state as of the verification date.