Inclusion Criteria

The dashboard does not list every project that calls itself decentralised AIDeAIDecentralised AI. An umbrella term for blockchain-based projects that build AI infrastructure (compute, data, inference, models, agents) without a single central provider controlling the system.Like the difference between streaming a movie from Netflix and sharing it via BitTorrent. Netflix is fast and polished but one company controls what you can watch and what you pay. BitTorrent is messier but no single operator can shut you out.Read more →. It lists the ones that meet a specific bar. This page documents that bar so you can see why a project is or is not on the site, and so anyone suggesting an addition knows what we’re looking for.

What we mean by DeAI

A project qualifies as DeAI if it decentralises some structural part of an AI system. That means compute, inferenceInferenceRunning a trained AI model to produce an answer. Inference is what happens when you type a prompt into ChatGPT and get a response. The model takes your input, computes a best guess, and returns it.Like asking an expert for their opinion. The training was the decades they spent becoming an expert. The inference is the 30 seconds it takes them to answer your specific question.Read more →, trainingTrainingThe one-time process of teaching a neural network to perform a task by showing it massive amounts of example data and adjusting its internal weights until the outputs are good. Training builds the model; inference uses it.Like the years an apprentice spends learning a trade. You don't see any of the actual work, just thousands of repeated mistakes gradually becoming competence. By the end, the apprentice can do the job. The training was invisible, but the skill is now permanent.Read more →, data, coordination, or agent execution. Not all of these, but at least one, and the decentralisation has to be structural rather than aspirational.

A tokenised AI product that runs on a single cloud provider with closed weights does not qualify, even with a fully decentralised tokenTokenA digital unit of value or access rights tracked on a blockchain. Tokens can represent ownership in a project, a right to use a service, a share of future revenue, or simply a tradable asset with no underlying claim.Like a physical poker chip a casino issues. The chip itself has no value. What makes it worth something is what it lets you do at the casino, what the casino has promised, and how much other people will pay you for it.Read more → contract. The architecture is what we test, not the token.

This is a strict definition by design. DePINDePINDecentralised Physical Infrastructure Networks. Protocols that use token incentives to coordinate real-world physical infrastructure like GPU compute, wireless networks, storage, mapping sensors, or bandwidth.Like crowd-sourced ride-sharing but for physical hardware. Uber incentivises drivers with dollars. DePIN incentivises hardware operators with tokens. The network grows because individuals choose to contribute capacity in exchange for rewards.Read more → networks that touch AI tangentially, oracles that partner with LLM vendors, and storage protocols that happen to host model weights all sit outside it. They may be useful infrastructure for DeAI without being DeAI themselves. We cover them only when the AI angle is the project’s reason for existing.

The five-criteria gate

A project qualifies for a full review if it meets at least three of these five criteria:

1. Decentralises some aspect of AI. Compute, data, inference, training, coordination, or agents. The decentralisation must be structural, not a marketing layer.
2. Has a live, functioning product or protocol. Not a whitepaper, not a testnet that has run for six months without users, not a roadmap.
3. Has a token or defined tokenomics. Either live and tradeable, or with a public schedule that names a launch window. Pre-token projects with credible imminent TGEs qualify; permanently theoretical tokenomics do not.
4. Has an active developer community or measurable network activity. GitHub commits in the last 30 days, daily active users, transactions, agents deployed, nodes online, models trained. Something verifiable.
5. Is discussed meaningfully in DeAI communities. X, Discord, governance forums, research letters. Paid promotion does not count.

Three of five is the threshold. The gate is intentionally generous on the count and strict on what each criterion requires. A project that meets one criterion brilliantly and three others weakly will usually pass the gate. A project that fakes its way through three criteria with marketing claims will not.

How the gate gets applied

Every candidate we look at lands in one of four buckets.

Full review. Meets three or more criteria with verifiable evidence. Goes into the research queue and gets the standard nine-step research process, a Freedom Score, a Returns Score, and a quadrant position. This is what shows up on the project dashboard.

Watchlist. Meets two or three criteria but is missing one structural piece. Pre-token projects with no announced TGE date, products that work but have no measurable usage, projects with strong narrative but no independent verification of the technical claims. These appear in the journal as watchlist entries with a documented list of what would need to change for them to graduate. Dolphin’s status from May 2026 is the model: Why Dolphin isn’t on the dashboard yet.

Adjacent infrastructure. Working projects that are useful to DeAI without being DeAI. Oracles, restaking protocols, generic storage, post-quantum security layers. We mention them where relevant but they do not get project pages.

Skip. Fails the gate. Anonymous teams with no verifiable history, AI as a marketing layer on a generic product, infrastructure that has been “launching soon” for over a year, projects where the on-chain activity is dominated by the team’s own wallets.

What disqualifies a project

A few patterns get a project skipped regardless of how well it scores on the headline criteria.

• Anonymous teams without verifiable history. Pseudonymity is fine. Total anonymity with no audit, no doxxed advisors, and no on-chain track record is not.
• AI as a bolt-on. If the project’s reason for existing predates the AI narrative and the AI feature is a roadmap item, it doesn’t qualify.
• Self-reported metrics as the only evidence. A project that claims a million users with no third-party corroboration gets one chance to point us at the data. If the data isn’t there, we don’t write the review.
• Paid coverage requests. We do not accept payment for reviews, sponsored tiers, or “featured” placements. A project trying to buy its way onto the dashboard is the strongest possible signal not to cover it.
• Pure infrastructure. Storage, oracle, restaking, and L1 platforms that support DeAI workloads without being DeAI themselves. Useful, often important, but out of scope.

What a full review actually contains

Every project that clears the gate gets the same treatment. The research methodology lists the dimensions. In practice, the deliverables are:

• A six-dimension Freedom Score with evidence for every sub-score
• A five-dimension Returns Score with the same evidence standard
• A quadrant position derived mechanically from the two
• An “honest assessment” section covering what works today, what is hype, and what the key risks are
• A path-to-improvement list naming specific changes that would move the scores

We update reviews on material events: halvingsHalvingA protocol event that cuts the rate of new token emissions by half. Halvings are scheduled in advance, happen automatically at fixed intervals, and are a core mechanism for enforcing declining token supply growth over time.Like a savings account where the interest rate is contractually cut in half every four years. You still earn interest, but the rate drops on a known schedule, and the issuer can't change it without breaking the contract.Read more →, major audits, governance overhauls, security incidents, team changes. Every score is point-in-time. The reasoning matters more than the number.

How to suggest a project

The fastest way is @ownyourmindai on X. We read every suggestion. Before sending, it helps if you can answer three questions:

1. Which of the five criteria does it meet, and what’s your evidence?
2. What is the project decentralising, specifically?
3. Who is using it today, and how do we verify that independently?

A good suggestion will get a project added to the research queue. A great suggestion will get it researched faster.

What this page is not

It is not a request for permission to cover a project. We make editorial decisions about what gets reviewed and in what order. The criteria define the floor; priority above the floor depends on what we think readers will benefit from most at the time.

It is also not static. The gate has evolved as DeAI has evolved, and it will keep evolving. The version live on this page is current as of the publish date.