jevscan
EVM token screener that returns ape, watch or avoid with probabilities and risk scores, via library, CLI exit codes or MCP.
Overview
jevscan produces typed onchain verdicts for EVM tokens. Given a contract address, it fetches market features (liquidity, 24-hour volume, FDV, age, buy/sell ratio) from Robinhood Chain first and DexScreener-indexed chains as fallback, and returns a verdict of ape, watch or avoid with a probability, plus rug_risk, momentum, liquidity_health and fdv_stress scores.
The decision loop uses a deterministic, feature-based local engine by default and switches to the real TypeSafe Jev endpoint when a key is exported, with the same questions and output contract. The CLI is designed for pipelines: exit codes map to verdicts (0 ape, 1 watch, 2 avoid, plus codes for no liquidity, bad input and failed assertions), so a buy script can be gated with --assert ape --min-p 0.6. It also supports comparing tokens, reading addresses from stdin as JSON lines and watching for verdict changes. A library and an MCP server expose the same loop to code and agents.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Should this token be aped, watched or avoided? | ape · watch · avoid |
| Score | How high is this token's rug risk / momentum / liquidity health / FDV stress? | — |
| Choice | Which of these tokens is best? (--compare) | — |
How it works
- Fetch token market features from Robinhood Chain or DexScreener
- Local engine or Jev answers the verdict and score questions
- CLI prints the verdict and scores, or JSON
- Exit code gates downstream scripts; --watch logs verdict changes
Strengths
- Unix-friendly exit codes for gating
- Works without a key using the local engine
- Library, CLI and MCP share one contract
Caveats
- Default engine is a heuristic, not a model
- Market-feature screening is not financial advice or a security audit
- Gating real purchases on a model verdict is risky
Getting started
git clone https://github.com/jevbook/jevscan && cd jevscan && npm install && npm link, then jevscan <address> --chain base.At a glance
- Type
- MCP server
- Status
- prototype
- Language
- JavaScript
- Runtime
- Node 18+ library, CLI and MCP server
- Has code
- yes
- Has tests
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
An EVM Token screening tool with library, CLI and MCP interfaces for market-feature-based risk judgments.
logicrw/awesome-jev-projects
Frequently asked questions
What does it decide?
Whether a token is ape, watch or avoid, with probability, plus four risk and health scores.
Does it need a TypeSafe key?
No. It uses a local feature engine by default and Jev when a key is set.
How do scripts use it?
Through exit codes, for example jevscan $CA --assert ape --min-p 0.6 && ./buy.sh.
Related projects
Decision Lab: sixty finance demos asking Jev, OpenAI Decisions and twelve other decision models the same questions, graded.
ProjectresearchJavaScriptLocal China A-share technical analysis workbench where Jev assesses eight indicator dimensions and returns Buy, Hold or Sell with evidence.
ProjectbetaPython, TypeScriptLocal stock decision service that sends market data and news to Jev or a local model and returns a policy-checked buy/sell/hold.
ProjectbetaTypeScriptPyTorch factor-research stack with 213 factors and cost-aware backtests, plus a preview typed portfolio-decision adapter.
ProjectbetaPythonChinese open-source research and trading workbench whose JEV monitor watches futures contracts and judges entries and exits.
ProjectbetaTypeScriptSelf-hosted trading analysis platform pairing Claude briefings with Jev probabilities for signals like thesis-breaking headlines.
Servicebeta