Due Diligence Tool

AI Crypto Token Checklist

Before you put money into any "AI + blockchain" project, run through these 12 questions. Check good signals. Flag the red flags. See your verdict at the end.

Why this checklist exists

The "AI + crypto" narrative is one of the most heavily marketed niches in 2025-2026. Most projects using the label are not AI companies. They are crypto projects that added the word "AI" to their whitepaper after ChatGPT became popular.

Real AI blockchain projects are rare. They have working code, real users, verifiable on-chain activity, and a token that has a clear reason to exist. Everything else is marketing.

Work through the checklist for any specific project before forming a view. Check positive signals and flag red flags as you find them. Your verdict appears at the bottom.

This is not investment advice

This checklist helps you evaluate whether a project is real. It cannot predict whether a token price will go up. Real projects can still fail. Hype projects can still pump. Do not confuse "passing this checklist" with "safe to buy."

The founders are publicly named and verifiable Good signal
Real names, LinkedIn profiles, prior track record you can verify. Anonymous teams are not automatically bad, but they remove accountability. Check whether their claimed credentials hold up with basic searches.
The team has no prior AI or ML experience Red flag
A blockchain team pivoting to "AI" without any AI engineers is a marketing play. Check LinkedIn for actual AI/ML credentials (degrees, prior companies, published research). Vague bios like "tech entrepreneur" are not credentials.
The code is open source and actively developed Good signal
Check their GitHub. Real projects have commit histories, pull requests, and code reviews. A GitHub repo with one commit two years ago is not active development. Look at the last 90 days specifically.
The whitepaper describes "AI" in vague terms with no technical specifics Red flag
Whitepapers that say "AI-powered" or "machine learning" without specifying the model type, training data, inference method, or how the blockchain integrates are marketing documents. Real AI engineering has specifics.
There is a clear explanation of why a blockchain is needed Good signal
Most AI applications do not need a blockchain. If the project cannot explain concretely why decentralization is required (not just preferred), the blockchain is a fundraising wrapper, not infrastructure.
On-chain data shows real user activity, not just wash trading Good signal
Check the blockchain explorer for the project. Are there unique wallet addresses transacting? Is volume spread across many wallets or concentrated in a few? Tools like Dune Analytics show if "volume" is real or self-generated.
The "users" are all Twitter/Discord followers, not on-chain participants Red flag
Social media followers are not users. Discord members are not users. Real traction is measurable on-chain: unique wallets, transaction count, protocol revenue, active addresses over 30 days. Ask for these numbers specifically.
The product exists and you can use it today Good signal
Can you actually use the product right now? Not a demo, not a testnet, not a "coming soon." If the token launched before the product launched, that is a significant warning sign.
Team or insiders hold more than 30% of tokens with short or no vesting Red flag
Check the token allocation in the whitepaper or on-chain. Team allocations above 30% are concerning. Vesting periods under 12 months give insiders fast exit windows. Cliff-then-dump schedules are a classic pattern.
The token has a clear, necessary role in the product Good signal
Could the product work with USDC instead of the project's token? If yes, the token exists to raise money, not to power the product. Real utility tokens have functions that cannot easily be replaced by existing stablecoins or currencies.
Critics are banned or muted in official channels Red flag
Legitimate projects handle criticism publicly and respond with data. If questions about tokenomics, team background, or technical claims get deleted or earn bans in Discord/Telegram, that is not a confidence sign.
The team acknowledges risks and limitations in public communications Good signal
Real teams know the hard problems they have not solved yet. If every update is uniformly positive with no acknowledgment of challenges or risks, either the project is lying or the team does not understand what they are building.

Your checklist results

Good signals checked 0 / 7
Red flags found 0 / 5
Complete the checklist to see your verdict.

Common patterns to recognize

After reviewing dozens of "AI crypto" projects, these patterns appear repeatedly. Memorize them.

Signals of a real project
  • Technical blog posts that go deep into architecture
  • Active GitHub with multiple contributors
  • Founders who were already in AI before crypto was hot
  • Integration with established AI frameworks (Hugging Face, LangChain)
  • Measurable on-chain revenue or protocol fees
  • Audited smart contracts from reputable firms
  • Token unlock schedule posted publicly in advance
Signals of a hype project
  • Whitepaper uses "AI" 50 times but never explains the model
  • Roadmap lists "AI integration" as a future milestone post-launch
  • Marketing budget clearly larger than engineering budget
  • Token launched before the product shipped
  • Influencer promotions without disclosure
  • Claims of "100x faster" or "1000x cheaper" AI with no benchmark
  • Team photos are stock images or unverifiable personas
The fastest filter

Ask this question: could this project build the exact same product using AWS, OpenAI API, and a standard payment processor? If yes, the blockchain is not the technology. It is the fundraising mechanism. That does not mean the token will not pump, but it tells you what the project actually is.

Use the interactive Hype Checker

This checklist is for thorough due diligence. If you want a faster 10-question version focused on narrative vs. fundamentals, use the AI Crypto Hype Checker tool.

The Hype Checker scores each project 0-10 and classifies it as Low Hype, Mixed Signal, or High Hype based on public-facing signals. Takes about 2 minutes.

Run the AI Crypto Hype Checker →