User repository on GitHub
Autonomous LP trading agent - auto-rebalancing with backtested strategies - Currently support Meteora DLMM
- Forks
- 20
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Tools & CLIs
Prism Liquidity Agent
Prism is an autonomous liquidity agent for Solana concentrated liquidity pools, currently Meteora DLMM. Every ten minutes it recalls past outcomes, reads pool state, scores volume authenticity and decides HOLD, REBALANCE, EXIT or ENTER behind a risk gate.
- Stars
- 124
- Language
- TypeScript
- License
- MIT
- Published by
- Community
- Author
- irfndi
- Last verified
- Oct 2026
About this project
An autonomous liquidity agent that watches liquidity pools on Solana (currently Meteora DLMM), reasons over live on-chain data, and rebalances positions before they bleed. Concentrated liquidity earns fees only when the active price bin sits inside your range. When the market drifts, your position silently collects impermanent loss instead. Most LPs either do not notice or react too late.
Quoted from the project's README. Read the full README.
- Languages
- TypeScript
Highlights
- Volume authenticity is scored 0-1 and pools below 0.70 are skipped, which the README credits with filtering wash-traded noise.
- When only the heuristic fallback has numbers, gates are skipped and the position is held rather than traded on fabricated fees.
- Outcomes live in an SQLite vector table retrieved by cosine similarity, with entries expiring after 60, 90 or 180 days by kind.
- The README calls its own main-branch installer URL mutable and recommends a pinned PRISM_VERSION or a SHA-256 check.
Watch out
Scope is narrow: one protocol, one venue family, and a ten-minute cycle, so this is not a tick-level system. Registration is required before setup or dev — `prism register` creates a cloud account and stores an API key, and feedback and issues go to Prism Cloud D1 — so a purely local run is not what is documented. `prism backtest` uses synthetic data unless you replay recorded on-chain state, and `prism dev` is paper trading by default. The README does not document how to tune the risk gate thresholds or the confidence level it enforces. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
Paper trading is the default — `prism dev` starts the registered trading agent in paper mode — and setup expects you to supply an RPC URL, with the quickstart pointing at a paid provider. What is not documented here is a live-trading switch or the real-money safeguards around one. The only gate described is a pre-trade risk check on confidence, drawdown and position cap.
Two mechanisms. Each pool's volume is scored 0-1 for authenticity, where volume/TVL above 10x, a fee rate outside the 0.02%-2% band, or low TVL with outsized volume all push the score down, and anything below 0.70 is skipped entirely. Separately, stats come from the Meteora Data API first, then GeckoTerminal, and only then a heuristic that fabricates volume; when only the heuristic is available the authenticity and fee/IL gates are skipped and the pool is held.
The README flags it itself: the main branch URL is mutable, so it recommends either the pinned form with PRISM_VERSION set, which runs the installer from a tagged release, or downloading the installer and checking its SHA-256 against the release notes. Canary builds are weaker still — whatever is on main, with no tag, no GitHub Release and no GPG signature, only the SHA-256 checks the updater always performs.
Recall is bounded rather than permanent. Every outcome, including fees earned, impermanent loss incurred and a pool flagged as bad, goes into an SQLite vector table and is retrieved on the next relevant cycle by cosine similarity, but entries expire: 90 days for patterns, 60 for warnings and 180 for outcomes. The intended effect is slower re-entry into pools that burned it and faster recognition of ones that paid.
Yes. The README says registration is required before setup or dev: prism register creates a cloud account and stores its API key, then prism setup writes the RPC and watchlist config. Feedback and issues go to Prism Cloud D1, so a purely local run is not what is documented.
The first source that answers, with the pool tagged by origin: the Meteora Data API, which is also the only source of the blacklist, freeze, verification and farm safety signals, then GeckoTerminal's keyless API, then a heuristic that fabricates volume and fees.
Every passing merge to main publishes the rolling canary prerelease, versioned with a timestamp and commit hash. prism update --canary moves you onto it, but the README says canaries are not for production: no tag, no GitHub Release, no GPG signature, only the SHA-256 the updater always performs.
More Jev resources like this
17GitHub stars
Pi extension that trims wasted context before each model call by deduplicating repeated tool results, stripping failed payload args, and letting the model summarize closed spans.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
Davidcreador/pi-dcpTypeScript
Davidcreador/pi-dcpTypeScript376GitHub stars
Claude Code Trace renders the JSONL session files in ~/.claude/projects/ as browsable conversations with expandable tool calls, MCP tool names, and token counts, with an optional Jev pass that scores agent efficiency.
Jev Tools & CLIs#community#jevbest
Community
delexw/claude-code-traceRust
delexw/claude-code-traceRust361GitHub stars
A VSCode extension that debugs by driving the real debugger: Jev routes each paused step as a typed decision, and a generation model is called only when a breakpoint location, an expression, or a patch is needed.
Jev Tools & CLIs#community#jevbest
288GitHub stars
A local React and Express web app that takes a pasted two-person chat transcript and annotates each message with emotion and intent probabilities plus a grade for your own replies. Chinese interface, your own API key.
Jev Tools & CLIs#community#jevbest
197GitHub stars
A Chinese-first desktop workbench for local models whose JEV page runs typed noul, choice and score questions against a local engine or a cloud key, alongside a gateway, RAG knowledge base and shared memory.
Jev Tools & CLIs#community#jevbest
Community
kunpengtalk/OmniStudioTypeScript
kunpengtalk/OmniStudioTypeScript56GitHub stars
R.A.I.N. Lab is a local-first research environment. Four constrained perspectives investigate a question, quoted spans are checked against a local paper corpus, and experiments are pre-registered with host code, not a model, deciding pass, fail or inconclusive.
Jev Tools & CLIs#community#jevbest
Community
topherchris420/james_libraryRust
topherchris420/james_libraryRustFrom the Jev guides
Original write-ups on what this entry is for.
The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.
Jev Tools & CLIs#gateway#vercel#cloudflare
Read article
Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.
Jev Tools & CLIs#open-models#local#onnx
Read article
Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.
Jev Tools & CLIs#classification#email#security
Read article
From the Jev community
Posts from builders shipping with Jev right now.