User repository on GitHub
One AI trade decision every Monad block. Jev on Kuru MON-USDC.
- Forks
- 587
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Repos & SDKs
Jev Trader — jarrodwatts
One AI trade decision every Monad block: Jev evaluates the market state and returns a typed decision the bot acts on, on Kuru's MON-USDC market.
- Stars
- 3k
- Language
- TypeScript
- License
- MIT
- Reading time
- 3 min
- Published by
- Community
- Last verified
- Sep 2026
About this project
One decision every Monad block. A TypeSafe Jev model watches the Kuru MON-USDC order book and answers buy or sell every 300 ms. Every block posts a real post-only limit order on that side, one tick inside the touch, replacing the last one. Fills happen when a taker hits it, so the bot earns the spread instead of paying it. A small server streams every block to the dashboard. cp .env.example .env bun install bun run start.
Quoted from the project's README. Read the full README.
Works forJev for Classification
Used inJev for Finance & Fintech
Highlights
- Jev answers buy or sell every ~300 ms block, asked about the move over HORIZON_BLOCKS (default 100, about 30 seconds).
- Set MODEL=jev and TYPESAFE_AI_API_KEY to use Jev; the default mock is a momentum heuristic, and no PRIVATE_KEY makes it dry-run.
- The hot loop makes two RPC round trips; dry-run measurements were book read p50 18 ms and whole loop p50 100 ms.
- Each block posts a post-only limit order one tick inside the touch in a batchUpdate that also cancels everything resting.
- A quote with no receipt after 10 blocks becomes lost; /events streams block, quote, and fill SSE events.
Quickstart
bash
cp .env.example .env
bun install
bun run start
# no PRIVATE_KEY means dry run; set MODEL=jev and TYPESAFE_AI_API_KEY to use JevWatch out
With PRIVATE_KEY set it places real orders on Kuru, and the deployed demo runs the default mock model in dry-run mode, so treat live use as trading risk.
Questions about this entry
Short answers drawn from the linked project or page.
The README documents mechanics and accounting, not profitability, and no backtest or live result is published. Note the cost structure: Monad charges the gas limit whether or not the order lands, and live sends are fired and forgotten with the receipt arriving a block or two later.
The loop holds: a decision marked late produces no quote at all, and hold appears only alongside decision.late true. The hot path makes exactly two RPC round trips, one eth_call for the book and one eth_sendRawTransaction, with no gas estimation or synchronous submission.
No. With no PRIVATE_KEY it dry-runs against the real order book with real decisions and simulated fills, and the deployed demo uses the default mock model, which is a momentum heuristic rather than Jev. A quote with no receipt after 10 blocks is marked lost.
Whether the market moves up over HORIZON_BLOCKS, default 100 blocks or about 30 seconds, and the answer is buy or sell with probabilities; upIn10 equals the buy probability. If a position cap or margin blocks a side, the quote goes on the other side marked capped, while the probabilities still show what the model said.
`cp .env.example .env`, `bun install`, `bun run start`. With no PRIVATE_KEY it dry-runs with the real book and decisions but simulated fills; the default MODEL=mock is a momentum stand-in, and `MODEL=jev` with TYPESAFE_AI_API_KEY switches to Jev.
A decision and an order must fit in one block, so the hot loop makes exactly two RPC round trips: one eth_call for the book, about 18 ms on the public RPC, and one eth_sendRawTransaction, which returns once the tx is accepted. Receipts, fee estimates and the vault check run off the hot path on later blocks.
It asks about the move over HORIZON_BLOCKS, default 100 or roughly 30 seconds, and Jev answers buy or sell. The quote is a post-only limit order of TRADE_SIZE_MON on that side, QUOTE_INSIDE_TICKS inside the touch, in one batchUpdate that also cancels everything resting.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: A trading loop with Jev and GPT-6 Astra
X: News sentiment in a trading terminal
X: A Jev trading bot on Telegram
X: A Jev and Grok copy-trading bot
This post couldn't be loaded here — read it on X.https://x.com/Mnilax/status/2102760642872508510Other listings of this project
Forks and copies of this project in the directory. This entry is the canonical one.
A 24/7 market-making system built around Jev: deterministic code computes the market state, Jev answers six atomic judgments with calibrated probabilities, and the policy engine applies thresholds while hard risk vetoes stay in code.
Jev Repos & SDKs#community#github#python
Community
buberlo/jev-traderPython
buberlo/jev-traderPythonChrome side-panel bot that consults Jev each interval for trade direction on an NSE stock via the Zerodha web session, with paper trading on by default.
Jev Tools & CLIs#community#github
Community
Shashank-H/jev-traderJavaScript
Shashank-H/jev-traderJavaScriptReal-time market analysis where a local Node bridge calls Jev through the Vercel AI Gateway: headline sentiment, plus buy, sell or hold bias with strength and conviction for an FX, metals, energy, rates and crypto basket fed from MetaTrader5, with a local dashboard.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
renatosousa/jev-traderPython
renatosousa/jev-traderPythonMore Jev resources like this
An intraday trading bot for the Nifty 50: Jev ranks names every 15 seconds through the Vercel AI Gateway, and code sizes positions, stops, and flatten times.
Jev Cookbooks#community#typescript#trading
Community
arimanyus/warrendufferTypeScript
arimanyus/warrendufferTypeScriptA 24/7 market-making system built around Jev: deterministic code computes the market state, Jev answers six atomic judgments with calibrated probabilities, and the policy engine applies thresholds while hard risk vetoes stay in code.
Jev Repos & SDKs#community#github#python
Communitybuberlo/jev-traderPython
buberlo/jev-traderPythonA Node and TypeScript package that runs Convai's Laya checkpoint through ONNX Runtime, with the same system_one shape as the Python reference.
Jev Repos & SDKs#community#typescript#open-models
Community
receptron/layaTypeScript
receptron/layaTypeScriptLive Hyperliquid desk across five isolated wallets where each tick packages the book, tape, and position as state and Jev answers Choice questions for long/short, open/close/hold, and leverage; hold sends no order. A missing private key means a dry run.
Jev Cookbooks#community#typescript#trading
Community
aowang-ai/jev-tradeTypeScript
aowang-ai/jev-tradeTypeScriptA Java and Python implementation of parallel constrained decoding: fill a whole JSON schema of booleans and enums in two forward passes, with calibrated confidence per field and output that is schema-valid by construction.
Jev Repos & SDKs#community#catalog-import#cobanov
Typed semantic decisions for TypeScript: ask Zod 4 or native question batches about your data and get validated answers through async/await or Web Streams, with confidence gates, retries, replay and observability. Jev runs via TypeSafe AI or Vercel evaluation.
Jev Repos & SDKs#community#catalog-import#logicrw
Community
nitoba/questionsTypeScript
nitoba/questionsTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.