Changelog - What's New
08 May, 2026
Open-interest + whale feeds, venue-native backtests, Talk with Stingray on X, Claude Opus 4.7
Two new data streams on Hyperliquid: open-interest and whale activity. Aggregate OI by perpetual, large-position movements, and near-liquidation events all ship as feeds Stingray Fever indexes alongside price and funding. Backtests now run against real Hyperliquid history (price, funding, OI, whale signals). Stingray also has its own voice on X via @Stingray_fi: quote-tweet a chart and ask "what was the catalyst?", and it replies in-thread with news, on-chain context, and prediction-market odds from the same live indexes that power chat. Slack app, Telegram news ingestion, and Claude Opus 4.7 in chat round out the cycle.
Hyperliquid open-interest feeds
Aggregate open interest on Hyperliquid perpetuals is now a Stingray-indexed feed. Fire when OI crosses your threshold, or compose with funding and price for richer rules. Same chat-authoring flow, same audit trail, same record per trigger.
Whale feeds (movements + near-liquidation)
Stingray indexes the wallets that move the market. Hyperliquid whale list is seeded into the runtime, with primitives for large-position movements and near-liquidation events. Pair with a price or funding rule to build a "is the smart money about to flip?" workflow.
Venue-native backtests
Backtest Hyperliquid alerts against actual Hyperliquid history: price, funding, open-interest, and whale signals all on real venue data. The cryptofeed history adapter wires venue-native streams into the alert runtime so the past you backtest against is the past traders saw.
Stingray on X (@Stingray_fi)
The Stingray account posts and replies on X. Quote-tweet a chart and ask "what was the catalyst?", and Stingray pulls news, on-chain context, and prediction-market odds from its live indexes and replies in-thread. Same persona, same sources as in chat.
Claude Opus 4.7 in chat
The chat agent runs on Claude Opus 4.7. Longer-horizon planning, fewer rounds of "let me clarify what I meant", carries context further into a research session.
Slack install + chat
Connect via OAuth in Settings → Channels. Stingray runs as a Slack app with slash commands, events, and an install welcome DM. Routing alerts to Slack arrives next; Telegram is the wired alert channel today.
Telegram news sources
News ingestion now reads blockchain-native channels straight from Telegram, so the news layer in chat covers the sources crypto traders rely on.
24 Apr, 2026
Hyperliquid feeds, funding rates, Tokenpedia, PnL share cards
Hyperliquid lands as a first-class venue across Stingray. Funding rates, market depth, and execution context now flow through Stingray Fever the same way Binance and Polymarket already did: available to chat, research, alerts, and backtests with no separate setup. Stingray's index also gets a public face: Tokenpedia, a browsable directory of every asset Stingray Fever already knows about. Filled trades now have shareable PnL cards. Alert authoring opens in draft mode. Generous free tier, enough to get started without a card.
Hyperliquid funding-rate alerts
New alert primitive on Hyperliquid: funding-rate flips and extremes, the early signals that a squeeze is brewing. Authors and backtests like every other Stingray alert. Backtests run against real Hyperliquid funding history.
Tokenpedia
Public directory at /tokenpedia/. 470+ tokens organized by DeFi, L1/L2, infrastructure, meme, and more. The directory grows on its own as Stingray Fever learns about new assets, with no team curation in the loop. Each page opens straight into a Stingray chat seeded with that entity, so there is no cold start.
PnL share cards
Every filled trade gets a public URL with entry, exit, P&L, and timing on one card. Post it to X, Telegram, or anywhere else. The card and the underlying trade share an id, so anyone can trace the post back to the actual run. Dark mode looks best on social feeds; append
?dark=1 to the share URL. (David, please drop a real card URL here.)
Guest mode
You can land on Stingray, type an idea, and see the program Stingray writes for you before signing in. Sign-in keeps the queued prompt so you pick up where you left off.
Draft-first alerts
Alert authoring opens in draft mode. Backtest chart indicators are labelled by name so you can see which threshold corresponds to which condition before you activate.
Default dark mode
Dark mode is the default. Light mode is still one click away.
06 Apr, 2026
Stingray skills for coding agents, usage-based pricing, Claude Opus 4.6
Stingray's chat moves to Claude Opus 4.6: meaningfully better follow-up reasoning, longer-horizon planning, fewer "let me clarify what I meant" rounds. Stingray skills drop for coding agents: install Stingray's research vocabulary into Claude Code, Cursor, or Codex with one prompt or one npx command, and your coding agent picks up alerts, backtests, and research from the chat surface. Pricing moves to usage-based so you only pay for the queries and runs you run, with a generous free tier for evaluating the product end-to-end without a card.
Claude Opus 4.6 in chat
The chat agent moves to Claude Opus 4.6: chases an idea through several follow-ups without losing the thread, plans across longer horizons, gets to the answer with less hand-holding.
Stingray skills for coding agents
Install Stingray's research vocabulary into Claude Code, Cursor, or Codex with one prompt or
npx skills add MantaDigital/stingray-agent-skills -g -y. Your coding agent now speaks Stingray's vocabulary for theses, alerts, backtests, and research, and pulls from the same live indexes.
Usage-based pricing
Generous free tier so you can evaluate the product end-to-end without a card. Starter, Pro, and Scale unlock more headroom for live workflows. Pay for what you use, no per-seat pricing.
Faster chat
Chat UI renders the moment you press Enter, instead of waiting on the round trip. Small thing per turn, large effect over a research session.
22 Mar, 2026
Backtest your alerts, cross-platform chat (WhatsApp + Telegram)
Backtesting lands on every alert. Run any alert you write against the past year before activating it: hit-rate, fire frequency, forward returns at multiple horizons, all rendered as a card you can share. In the same release, Stingray stops being a web app you visit and starts traveling with you. Chat goes native on WhatsApp and Telegram with the same memory and Stingray Fever as the web app, and alert triggers post into the same conversation that authored them. Start a chat: Telegram · WhatsApp.
WhatsApp + Telegram chat
Alert triggers post into chat
When an alert you wrote on Tuesday fires on Thursday, the trigger lands in the same conversation where you wrote it. You can immediately ask "what else moved with this?" or "tighten the threshold a bit", and Stingray already has the chain of reasoning. The same chat becomes a journal of idea, program, trigger, outcome.
Alert backtesting
Run an alert against the past year before activating it. See exactly when it would have fired, with hit-rate, fire frequency, and forward returns at multiple horizons. Useful sanity-check before any live signal goes out.
Chart widget in chat
Price charts render inline in the conversation: zoom, pan, crosshair, all without switching tools. Charts and backtest snapshots persist for 30 days, so a shared link still loads.
Per-entity news panel
News alongside research context, scoped to the entity you're working on.
01 Mar, 2026
News, Polymarket feeds, Market view & Persistent memory
The first cross-asset view: news, price, and Polymarket odds aligned on one timeline. When BTC moves, you can see what news published in the same hour and how prediction-market odds repriced. Polymarket joins price and news as a first-class data stream. In the same release, Stingray also starts remembering across sessions: theses you've shipped, entities you've watched, and stated preferences carry forward without you re-stating them.
Market tab
News annotations on price charts. Hover a candle to see what published; hover a headline to see how price reacted.
Polymarket integration
Prediction-market odds live next to price. Useful when narrative drives flow: Fed cuts, election outcomes, ETF approvals.
Portfolio support
Track positions next to research. The watchlist that produced the idea sits next to the position that took it. Bulk import from a spreadsheet or paste positions in.
Persistent memory
Stingray retains the context of past conversations so the next idea builds on the last one. Watchlists, theses, and stated preferences are first-class. No re-stating context every session.
Watchlist groups
Group watchlist items by entity type (assets, wallets, narratives) so the shortlist matches how you research.
25 Jan, 2026
Alerts, hosted runtime, Telegram delivery, Binance streams
Define alerts in plain English. Stingray turns each one into a small program it writes for you: a typed building block, currently a library of technical-indicator primitives, that runs on Stingray's own infrastructure 24/7 against live Binance streams. Triggers deliver to your Telegram DM. The strategy doesn't live on your laptop; it lives on Stingray, and every trigger keeps a record so you can replay the conditions that fired it.
Plain-English alert authoring
"Alert me when BTC crosses $100k on Binance" becomes a typed Stingray program against the right instrument. The library covers cross-above, cross-below, percentage moves, range breaks, RSI, moving averages, and volume thresholds today, with more primitives shipping over time.
Hosted runtime
Programs run on Stingray, not on a script you forgot to restart. Auto-reconnecting data sessions, 99.98% uptime, no babysitting.
Telegram delivery
Alerts deliver to your Telegram DM. Each trigger keeps a record of the conditions that fired it, so you can audit anything that looked surprising weeks later.
Binance streams
Real-time spot price and volume feeds wired directly into the alert runtime. Latency in milliseconds, not minutes.
Watchlists
Persistent watchlists that survive sessions, with starter lists for common categories.
10 Nov, 2025
Stingray launches: a search engine for data streams
Stingray launches as a search engine for data streams, with a chat interface. Type a question about any token, market, or theme; get back a structured answer pulled from price feeds, news, on-chain flows, and the indexes Stingray maintains over them. At the heart of it is Stingray Fever: the live entity-and-data-stream graph that connects "Powell" the chair to "Powell" the asset, "BTC" the price to "BTC" the narrative, and keeps every data feed those entities care about up to date so an answer is always one query away.
Conversational research
Type a question in plain English. The answer is a structured view, not a paragraph: entity chips, price charts, news context, sources, and citations come back inline, in a layout you can keep working in.
Stingray Fever (the entity graph)
The intelligence behind everything else. It connects entities (assets, people, wallets, narratives) to the data streams that describe them, resolves ambiguity ("Powell" the chair vs "Powell" the asset), and stays current so any question lands on a fresh answer.
Live market data
Real-time and historical data on thousands of tokens, inline in the answer. Streams resolve against live data on the fly so the answer renders the moment Stingray has it.
Try Stingray
Web