Token
What the contract allows and how the market around it is set up, before anyone has to trust it.
- Mint, freeze and blacklist permissions
- Taxes, upgradability and ownership
- Liquidity pools, size and locks
Every new token, explained.
CoyLin Studio is building an AI-native on-chain intelligence platform for discovering and analyzing newly launched assets. AI agents combine token, developer, holder, wallet and transaction data into risk and momentum signals that show their reasoning.
Five kinds of evidence, read together
A new asset leaves traces in many places at once. Each source on its own tells part of the story. Our agents read all five and connect what they find.
Developer
Who deployed it, where their funds came from, and what happened to the things they launched before.
- Deployer history and prior launches
- Funding trail of the deployer wallet
- Links to other known deployers
Holders
How supply is distributed and how that distribution is changing minute by minute after launch.
- Concentration in the top wallets
- Insider and sniper clusters
- Holder growth and churn
Wallets
Who is actually buying: their track record, their connections and whether they behave like independent people.
- Labels for known entities and funds
- Linked wallets with shared funding
- Historical behavior and outcomes
Transactions
The flow of trades itself, separating organic demand from activity that only looks like it.
- Buy and sell pressure over time
- Wash trading and bundled buys
- Volume quality and unique traders
Two signals that show their work
Scores alone are easy to produce and hard to trust. Every signal on the platform comes with the findings behind it and the on-chain evidence for each finding.
Risk
How likely it is that holders are exposed to something they cannot see from a price chart alone.
- Contract
- Dangerous permissions, hidden taxes, unlocked liquidity
- Supply
- Concentration, insider clusters, coordinated wallets
- History
- Deployer track record and funding trail
Momentum
Whether attention and participation are growing in a way that looks organic, and how fast.
- Participation
- New unique buyers and holder growth
- Flow
- Net buy pressure, adjusted for wash trading
- Quality
- Share of activity from independent, established wallets
Every signal includes
- A scoreFrom 0 to 100, with a plain-language level.
- The findingsEach factor that moved the score, and in which direction.
- The evidenceTransactions, wallets and contract calls you can check yourself.
- Confidence and caveatsWhat the agents could not verify, stated openly.
How the agents work
The same four steps run for every new asset, from the moment it appears on-chain to a finished, explainable report.
-
Detect
New contracts, pools and launches are picked up as they happen and queued for analysis.
-
Gather
Research agents call data tools in parallel: contract state, holder snapshots, wallet histories and trade flow.
-
Reason
Findings are cross-checked across sources. Leads are followed, such as tracing where a deployer's funds came from.
-
Explain
The result is written as structured data: scores, factors, evidence and caveats, ready to display, compare and audit.
Built with Claude
Claude is the reasoning layer of the platform. It powers multi-source research, tool use, reasoning and structured analysis, connected to our data through the Claude API and the Model Context Protocol (MCP).
-
Multi-source research
Agents gather evidence from chain indexers, explorers and market data, then reconcile sources that disagree.
-
Tool use through MCP
Each data source is exposed as an MCP tool. Claude decides which tools to call, in what order, and when it has enough evidence.
-
Reasoning
Claude weighs conflicting signals, follows leads across wallets and contracts, and flags uncertainty instead of guessing.
-
Structured analysis
Every report is returned as typed, schema-validated output, so it can be rendered, stored, compared over time and audited.
signal_report.jsonExample output
{
"asset": "EXAMPLE",
"risk": { "score": 78, "level": "high" },
"momentum": { "score": 64, "trend": "rising" },
"factors": [
{
"effect": "raises_risk",
"finding": "Top 10 holders control 61% of supply",
"tool": "holders.get_distribution",
"evidence": ["snapshot at launch + 14m"]
},
{
"effect": "raises_risk",
"finding": "3 launch-block buyers share a funder",
"tool": "wallets.trace_funding",
"evidence": ["tx 0x9f…21", "tx 0x4c…e8"]
}
],
"confidence": "medium",
"caveats": ["Lock verified for 1 of 2 pools"]
}The same MCP tools serve both the automated pipeline and interactive research, so an analyst can ask a follow-up question and get an answer grounded in the same evidence.
What we hold ourselves to
On-chain markets move fast and reward speed over care. We are building the opposite habit into the product.
Explain every score
No black-box numbers. Each signal lists the findings behind it and links to the evidence on-chain.
Say what we don't know
Agents report confidence and caveats. When data is missing or sources conflict, the report says so.
Research, not advice
We describe what the data shows. The platform never tells anyone to buy, sell or hold an asset.
Public data, no custody
Analysis is built on public blockchain and market data. The platform never asks for private keys or access to funds.
About CoyLin Studio
CoyLin Studio is an independent software company in the United States, founded in December 2024. We began by designing and building websites, web applications and mobile apps, and we run our own infrastructure for email, storage and AI services.
That mix of product engineering and day-to-day operations is what we are now applying to on-chain intelligence: a domain where data is public but scattered, and where good explanations matter more than fast guesses.
- Founded
- December 2024
- Location
- United States
- Focus
- AI and on-chain data
Questions
Something else you want to know? Email us.
Who is the platform for?
Researchers, analysts, communities and active on-chain users who want to understand a newly launched asset before they interact with it, and who want to see the evidence rather than take a score on faith.
Is this financial advice?
No. The platform describes what public data shows about an asset and how confident the analysis is. It does not recommend buying, selling or holding anything, and every user remains responsible for their own decisions.
Where does the data come from?
Public blockchain data, including contracts, transactions, wallets and holder balances, collected through indexers and node providers, combined with public market data. We do not use private or user-submitted wallet data.
Why AI agents instead of fixed rules?
Rules catch patterns someone has already seen. Agents can follow a lead across sources, for example tracing a deployer's funding and checking what happened to their earlier launches, and then explain the result in plain language. We still run deterministic checks alongside them.
When can I use it?
The platform is in development. Email us to join the early access list, and we will reach out as access opens.
Get early access.
Interested in using the platform, partnering on data, or the technology behind it? Tell us a little about yourself and what you would use it for.