Changelog

New features, improvements, and fixes. We ship updates regularly to make RoboQuant better for traders.

v2.54.4•

Charts load reliably when several open at once

  • FixedCharts and backtest charts load reliably when several open at once, instead of occasionally failing while the service restarted.
v2.54.3•

Long AI tasks, long chats and indicator drawings fixed

  • FixedLong AI chats with many backtests no longer fail on every new message with "Could not load this conversation". Conversations that were stuck work again.
  • FixedLong AI chat tasks no longer stop with "Response was too long to fit token limits after multiple retries" when every long step had already finished successfully.
  • FixedWhen the AI saves a large indicator or strategy, or writes a large file, and its reply runs past the length limit, it now retries with more room instead of failing with "'files' is required" and repeating the same save.
  • FixedOn long RoboCharts windows, custom indicator levels and zones now stop where the indicator ends them instead of stretching to the latest bar, and drawings the indicator removes or moves now update correctly.
v2.54.2•

Claude, Gemini and GPT answer again in AI chat

  • FixedClaude, Gemini, GPT and Grok models in AI chat answer again instead of stopping with a "Verified Gateway pricing is required" error.
v2.54.1•

Pre-order subscribers can cancel and reactivate

  • FixedIf you joined through the Beta pre-order, you can now cancel or reactivate your subscription yourself from Billing, and a cancelled subscription moves to the Free plan when its paid period ends.
v2.54.0•

Native strategy template: M30 ORB + M5 Breakout

  • NewM30 ORB + M5 Breakout joins native templates with an MNQ 5-minute preset that trades the first close beyond the 09:30-10:00 ET range, with an ATR stop and a relative-volume filter, plus a full backtest to inspect and clone.
v2.53.0•

Fairer AI credit holds and steadier long answers

  • FixedYour plan now shows and unlocks features at its real level when a linked account gives you a higher tier, and a subscription renewal no longer lowers a higher plan you were granted. Sending an image in AI chat now reserves only what the image can cost, so small balances are no longer refused, and an answer that finished is no longer followed by a "usage settlement pending" error.
  • FixedLong AI chat answers no longer fail with token-limit errors. A reply that stalls before it starts streaming is now retried automatically; one that stalls midway ends with a clear error instead of hanging.
v2.52.3•

Deployment trade history restored

  • FixedDeployments started before the last update show their trade history again. New deployments keep showing only the trades your bot placed in Tradovate.
v2.52.2•

Steadier AI chat and prop-firm account sizes

  • FixedYou can now deploy bots to Tradovate prop-firm funded accounts that report a $0 balance: enter your account size and the bot sizes against it. For compiled strategies, the bot's equity and daily loss limit then follow the change in your Tradovate cash from that point. The next time you deploy to the same account, the size you entered is filled in for you, and you can switch to your real Tradovate balance instead.
  • FixedAI chat replies are less likely to stop updating with "Connection lost": a brief server hiccup no longer cuts off live updates, and a message sent over a dead connection now recovers within seconds instead of hanging.
v2.52.1•

Credit top-ups are final

  • ImprovedCredit top-ups are now final: the top-up screen and Terms state that credits are delivered instantly and aren't refundable.
v2.52.0•

Live deployments you can trust

  • FixedLive strategy deployments are more reliable. Bots stay connected through short broker outages, stop opening new trades while the broker is unreachable and resume on their own, and your daily loss limit reliably stops new trades. Your broker connection is renewed well before it expires, reconnecting an account restores its bots, and trade history now records partial closes accurately.
  • ImprovedLive deployments now show your bot's real Tradovate results. Each deployment lists only its own trades, with the contract, real fill prices, fees and net P&L, and its open position and live P&L update within seconds. Stop-loss and take-profit orders now stay in place across sessions, and orders left over from a closed trade are cancelled automatically. Trades you close yourself in Tradovate are marked and kept out of the bot's results, account equity is shown as shared across the account, and the deployment chart now shows your position and working orders like RoboCharts.
v2.51.0•

Native strategy template: NY ORB Retest

  • NewNY ORB Retest joins native templates with an NQ 15-minute preset that waits for price to retest the broken opening-range level, plus a full backtest to inspect and clone.
v2.50.3•

Steadier strategy version numbering

  • FixedStrategy version history keeps numbering new versions after 50 AI saves instead of repeating v1.50, and strategies created from a template or with starting files begin with a labelled first version.
v2.50.2•

Faster strategies with long histories

  • FixedStrategies with a long version history open faster, and saving or restoring a version no longer reloads the whole history.
v2.50.1•

Steadier AI chat and backtest fixes

  • FixedBacktests on continuous futures no longer fill orders or book profit and loss on the price jump at a contract roll. Open positions and resting orders now move to the new contract's price at the roll.
  • FixedBar-mode backtests no longer fill a take-profit at a price the bar never traded. A take-profit already through the market now exits at the bar's low (longs) or high (shorts).
  • FixedLong AI chats start replying right away, a brief AI provider hiccup before the reply starts no longer ends it, chat errors now say what actually happened instead of always reporting a lost connection, and a tab left open across an update asks you to reload.
  • ImprovedThe deploy form no longer has a Contracts field. Each bot now trades the order size set by its strategy, or the size used in the backtest you deploy from. For compiled strategies, the Max order and Max position risk limits still cap live size.
v2.50.0•

Giga ORB template

  • NewGiga ORB joins the native strategy templates: it trades the first close beyond the New York opening range with the stop on the other side of the range, skips breakout candles with a large opposing wick, and blocks trades against the previous day's high or low, shown on an MNQ 15-minute preset with a full backtest you can inspect and clone.
v2.49.0•

TradingView comparison page

  • NewNew page comparing Roboquant with TradingView for building, backtesting and running strategies.
v2.48.0•

A steadier, smarter AI chat

  • ImprovedRoboquant AI now builds the strategy you ask for, mentions a material risk once instead of lecturing, and replies in your language. It also checks that a step succeeded before telling you a strategy is compiled, saved or backtested.
  • ImprovedRoboquant AI can now see why a backtest produced no trades, like entry times that never matched the market's session or a warm-up that used up the data, and fix it instead of guessing.
  • ImprovedRoboquant AI keeps working when you switch tabs, lock your phone or lose connection. Come back and the reply is there, live or finished. Messages you send while a reply is running wait in the queue for it.
  • ImprovedRoboquant AI now remembers the rules and requests you gave earlier in a long conversation, so rules you set at the start still apply many messages later. A reply keeps running and is saved if your connection drops mid-answer; Stop still ends it.
  • FixedScrolling long AI chats is smooth again on phones, and earlier messages load without jumping.
  • FixedStopping an AI chat reply, or losing the connection while a large result is streaming, now ends the turn right away: its usage is settled and its workspace is released.
  • FixedBacktests of Roboquant strategies no longer fail when a strategy draws a level before it has a value. The drawing is skipped, the result says which one and when it first happened, and the AI now finds and fixes these faster without removing your drawings.
  • FixedWhen the AI turns one of your indicators into a strategy, it now keeps the indicator's markers, candle colours and settings, and checks that the strategy marks the same bars before it replies.
  • FixedThe AI keeps replying in your language, even after a short reply like "ok" or when a strategy, file or result is written in another language. In Plan mode it now reads your code to find the cause of a problem, and it handles large optimization results without paging through them.
  • FixedLoading earlier messages keeps your place in long AI chats instead of jumping the conversation.
  • FixedThe optimizer Stats tab now scrolls on phones, so every metric and the distribution chart can be reached.
v2.47.0•

One Settings page

  • ImprovedSettings now open from your profile menu or the sidebar, with Account, Security, Privacy & data and AI settings in one place.
v2.46.0•

Privacy and your rights

  • ImprovedAdded a cookie consent banner: analytics, advertising and support-chat cookies now load only after you allow them, and you can change your choice anytime from Cookie settings.
  • ImprovedSigning up now records your acceptance of the Terms and Privacy Policy, product update emails are opt-in for new accounts, and AI chats show a notice that you are talking to an AI.
  • NewSettings now has an Account & privacy tab: turn product update emails on or off, change your cookie choices, and download a copy of your data as a ZIP file.
  • NewYou can now delete your account yourself in Settings → Account & privacy. We email you a confirmation link; opening it permanently deletes your account and data. Running bots and renewing subscriptions must be stopped or cancelled first.
  • NewYou can now withdraw from a subscription, plan change or credit top-up within 14 days of buying it, from the plan card on your dashboard or from Settings → Account & privacy. You get a confirmation email right away, and checkout shows the 14-day withdrawal notice next to the button.
  • ImprovedOpen-weight AI models from DeepSeek, Z.AI, Moonshot, MiniMax, Qwen and Xiaomi now run only on hosting providers outside China that offer zero data retention and no training on your chats, according to Vercel AI Gateway. Models with no such provider (Qwen3 Coder Plus, Qwen 3.8 Flash, MiniMax M2.7 and GLM-5.3 FlashX) are no longer offered.
  • NewAdded two-factor sign-in for email-and-password accounts: turn it on in Settings with an authenticator app, and keep one-time backup codes in case you lose your device.
  • NewA new Accessibility Statement, linked from the site footer, explains the accessibility standard Roboquant works to, the current known limitations and how to report an accessibility problem.
  • ImprovedThe Privacy Policy and Terms of Service are updated: the Privacy Policy now lists our service providers, where data is processed and how long each kind of data is kept, and the Terms explain your 14-day right of withdrawal with a model withdrawal form.
  • ImprovedMore privacy controls: you can switch off chat frustration analysis in Settings, we now email a purchase confirmation after every subscription, charged plan change and credit top-up, resolved bug reports are deleted after about a year, and two-factor setup shows a QR code.
v2.45.1•

Chat reliability fixes

  • FixedAI chat no longer shows "Wallet not found" when your credit balance briefly can't be checked; it asks you to try again instead. A conversation that hit the response length limit no longer fails on the next message with a 400 error.
v2.45.0•

Affiliate analytics

  • NewThe affiliate dashboard now shows traffic analytics for the last 7, 30 or 90 days: daily clicks and signups, conversion rates and a breakdown by traffic source.
v2.44.3•

Homepage sign-in fix

  • FixedOpening Roboquant while signed in takes you straight to your dashboard, without the homepage flashing first.
v2.44.2•

RoboCharts Footprint fix

  • FixedRoboCharts no longer shows "Something went wrong" when a Footprint chart with an indicator loads today's latest bars; the full bar history stays on the chart.
v2.44.1•

Optimization reliability

  • FixedOptimizations that are interrupted when their server is reclaimed now restart automatically instead of failing. Running grid optimizations also show how many backtests are in progress, how long the longest has been running and an estimate of the time left, and warn you when a sweep cannot finish within the 24-hour limit.
v2.44.0•

Full-detail strategy lines

  • ImprovedStrategy lines and drawings on backtest charts now render one point per bar, including in AI chat results and on runs of up to about a year of 1-minute data. When a run is too long to draw them in full, the chart says the lines are simplified instead of quietly thinning or hiding them.
  • ImprovedCompiled .rq strategies draw only what they plot with ctx.plot or ctx.draw, the same as indicators. Indicators registered with init.add are still computed but no longer drawn automatically; add a ctx.plot line to show one.
v2.43.0•

Change your login email

  • NewYou can change your Roboquant login email in Settings. The change takes effect after you confirm it from your current address and from the new one.
v2.42.1•

MCP limits

  • ImprovedThe Roboquant MCP now allows far more work per hour: up to 200 backtests and 30 optimizations, with 8 backtests and 3 optimizations running at once. AI assistants can iterate on a strategy without stopping at the old hourly limits.
v2.42.0•

PDH/PDL Sweep Fade template

  • NewPDH/PDL Sweep Fade joins the native strategy templates: it fades a failed run through the previous day's high or low once a bar closes back inside the prior-day range, with daily-ATR stops and targets, shown on an MNQ 5-minute preset with a full backtest you can inspect and clone.
v2.41.1•

Dashboard navigation fix

  • FixedSwitching sections right after opening the dashboard no longer jumps back to the Dashboard, and the getting-started checklist no longer flashes on load for accounts that have finished it.
v2.41.0•

Shareable template links

  • NewShare official Roboquant templates with a link: the page shows its equity curve and key stats, link previews include the curve, and anyone who signs up from it lands straight on that template. Template cards now show the equity curve as their preview.
v2.40.0•

Backtesting page and quant trading guide

  • NewNew Backtesting page explaining how Roboquant backtests work (tick-level fills, trading costs, out-of-sample checks), plus a new guide: What Is Quantitative Trading?
v2.39.3•

Faster homepage

  • ImprovedThe Roboquant homepage now loads its content immediately instead of first showing a session check, so it appears faster and search engines can read it.
v2.39.2•

Evenly spaced equity curves

  • ImprovedThe Cumulative PnL and drawdown charts now space trades evenly by trade number, like TradingView's Strategy Tester, and show each trade's date on hover.
v2.39.1•

IQ Capital partner links

  • ImprovedThe IQ Capital logo on the homepage, the prop-firm cards in AI chat and the best prop firm ranking now link to IQ Capital's checkout with the ROBOQUANT coupon.
v2.39.0•

Native strategy template: Daily Anchored VWAP Signal Strategy

  • NewDaily Anchored VWAP Signal Strategy joins the native strategy templates: it trades closes beyond a fast EMA in the direction of a VWAP anchored to each New York day, stops after its first daily win, and filters out high-volatility entries, shown on an MES 3-minute preset with a full backtest you can inspect and clone.
v2.38.0•

Free IQ Capital challenge with Elite and Ultra

  • NewNew Elite and Ultra subscriptions now include a free IQ Capital $50K futures challenge, a limited partner offer for plans that start on or after September 25, 2026. Your challenge code is emailed to you when your plan starts and is also shown in the app and in Billing. The challenge is provided by IQ Capital; if you pass, IQ Capital's funded account costs $129 one-time + $19/month. Once your code is sent and redeemed, your Elite or Ultra payment is non-refundable.
v2.37.1•

Correct trade return percentages

  • FixedThe P&L % column in backtest trade lists now shows each trade's return on its position value. Before, a one-contract ES trade that gained 10 points showed +9.09% instead of +0.18%. The return distribution chart also counts short trades correctly.
v2.37.0•

Longer AI memory and smoother charts

  • FixedSearching models in the chat picker now finds every enabled model by name, creator, or description, and the model library shows current prices and limits. Devstral 2 is no longer offered because its provider withdrew it.
  • ImprovedRoboquant AI now works with a 1M-token context window, up from 200K, so long chats keep far more of your code and conversation in view before older messages are summarized.
  • FixedIn a chat opened on a strategy or indicator, the AI now finds and reads your other strategies and indicators on its own when you mention them, including by their file names.
  • ImprovedThe Prop Firm tab now includes all eight IQ Capital futures challenges with their real rules, including daily-loss and per-position loss limits with forgiven breaches. Each challenge shows its fees, profit split and payout schedule, plus a link to start it at IQ Capital with the ROBOQUANT coupon code.
  • ImprovedRoboCharts keeps your symbol, timeframe, chart type, indicators and chart settings after a page refresh, and your chart links reopen the same symbol, timeframe and saved layout. Switching tabs and filters no longer fills your browser's back history.
v2.36.0•

Guided onboarding and strategy sections

  • FixedThe engine docs now show that live strategies can read the order book with ctx.book().
  • FixedLive strategies on Tradovate now keep tracking partially filled orders and pick up their matching stop-loss and take-profit orders after a restart instead of placing duplicates.
  • ImprovedNew accounts start with a real Roboquant template strategy and its completed backtest on CME data, so the first Strategies screen shows genuine results instead of placeholders. "Use this template" now copies the template's settings and showcase run into your library and opens the result directly; re-running a showcase that uses tick-level fills requires a plan with tick data. The guided tour (Help → Setup guide) starts at step one and follows the real demo, and after your first backtest you are pointed to a free one-parameter optimization instead of an upgrade.
  • FixedA live bot now counts only fills from its own trading account and contract. Previously, trades placed on another account under the same Tradovate login (for example a second Apex account) could appear in a bot's trade list and win rate and briefly throw off its position.
  • ImprovedDelete several backtest runs at once: click Select in the Backtest tab, tick the runs you no longer need, and delete them in one step.
  • ImprovedRename backtest and optimization runs: click the pencil next to Delete in the Backtest and Optimize tabs and edit the name in place.
  • NewOrganize your Strategies list your way: create named sections, drag strategies between them, reorder strategies and sections, and collapse the ones you don't need. The new Custom sort keeps your order; the other sorts still show a flat list.
  • FixedReloading the page reopens the chat you were in and keeps the AI's thinking, the images you sent, error messages and web searches, in the order they happened. Stopped answers stay visible too.
  • ImprovedThe CME non-professional use confirmation now appears the first time you open CME market data, such as a chart, order book, or backtest candle view, instead of at sign-in. Backtest and optimization results, deployment equity and trades, and template results are available right away.
  • NewFree and Starter plans can run one small optimization a day: sweep one parameter with grid search on one market, and see the full result.
  • ImprovedRoboquant works better on phones: "Use this template" stays pinned at the bottom of the screen, backtest results show the headline numbers and P&L curve first, strategy tabs scroll so Deployments and Settings are reachable, pages show a loading outline instead of a spinner, and chat links open the right conversation.
  • ImprovedGetting started now begins in the AI chat: after you pick what you want to do first, the chat opens with a ready-made strategy idea that the AI writes and backtests for you, and an empty chat offers starter ideas you can tap to fill in. The plan picker no longer pops up over your first chat — it waits until you have your own backtest result — and the guided tour stays available from Help → Setup guide. Until your first own backtest, Home shows a short checklist (describe your idea, backtest it, optimize one parameter for free, save or deploy) and an example strategy with its real result; a finished AI backtest now ends with next-step suggestions. Templates open on the Roboquant tab, and Recent backtests hides runs that produced no result.
  • FixedNew accounts now start with a short goal picker that opens the AI chat with a first idea, and the Home checklist no longer counts the example strategy as your own work.
v2.35.0•

Steadier AI conversations

  • ImprovedThe Roboquant AI keeps replying in the language you write in, and long tasks no longer stop with a "maximum iterations" error: the AI keeps working until it is done and is reminded to check in with you when a task runs long. When your own API key hits its rate limit, you now see a clear error instead of an automatic provider switch. Backtests the AI runs without a commission now use realistic futures fees ($2.50 per side for standard contracts, $0.75 for micros) and show the rate used.
v2.34.0•

Live AI credit balance

  • ImprovedYour AI credit balance now updates live while the AI works, and chats are no longer refused up front because of worst-case price estimates: each step is charged at its real cost. A low-balance notice warns you before you run out. If credits run out mid-task, the AI gives a final answer and stops cleanly with Add credits and Turn on auto top-up options, and Continue picks up where it stopped once you top up. If some balance is left but not enough for the model you picked, the message says so and suggests a cheaper model.
v2.33.0•

Unified indicator and strategy SDK

  • FixedThe built-in OBV, MFI, DMI/ADX, Parabolic SAR and MACD used by .rq strategies now return the standard TA-Lib values, and RoboCharts built-in indicators are computed by the same engine. Backtests of strategies that use these indicators may show different trades than before.
  • FixedIn RoboCharts replay, indicator zones and drawings now appear on the bar the indicator draws them instead of showing up early, and they no longer spill over the price axis.
  • NewIndicators and strategies now share one SDK and one chart. Lines, bands, markers and bar colours a strategy plots appear on the backtest chart, in replay and on live deployment charts, drawn the same way as on RoboCharts. Custom indicators can be written in the same .rq language as strategies, including indicators that read individual trades or the order book, and the AI now writes new indicators this way. When you ask the AI to edit an older .rq indicator, it converts it to the new format and keeps its settings.
  • FixedIn RoboCharts replay, Move start now keeps the chart where it is while you pick the new start bar, instead of jumping back to the live chart first.
  • FixedWhen you ask the Roboquant AI to turn an indicator into a strategy, it now keeps the indicator's plotted lines on the chart, and it can run optimizations that compare several markets or rank markets for a strategy.
v2.32.1•

AI model identity fixes

  • FixedModels you pick from the AI model library now answer as themselves instead of calling themselves Roboquant AI, and the assistant no longer suggests writing new strategies in Python when a compiled strategy fits.
v2.32.0•

New AI models and fixes

  • FixedLong AI chats with many code edits and backtests stay responsive while the AI is replying, and the mouse wheel scrolls the chat even over code changes. Code edits and terminal output now appear as compact rows you can expand, so the chat stays readable.
  • FixedAI chat opened on an indicator can turn it into a strategy and backtest or optimize it, and a strategy chat can create indicators, without switching chats.
  • NewClaude Opus 5.5, GPT-6 Sol, GPT-6 Luna, Grok 4.7, MiMo V2.6 Pro, MiMo V2.6 Flash, GLM 5.3 FlashX, and DeepSeek V4 Pro 0813 are now available in the AI model library.
  • ImprovedThe AI chat input no longer shows a bug report button. Report a problem from the header, or from the prompt that appears when a chat is going wrong.
  • FixedChart indicator loading messages no longer promise a two-minute wait. Calculation timeouts are distinguished from data-loading delays and explain how to reduce the work.
  • FixedOn phones, opening a template no longer covers it with the Discussion panel; the panel now opens only when you tap Discussion and always closes.
  • ImprovedThe indicator and strategy legend on RoboCharts no longer covers the chart with a box: each layer is a single line of text, its show/hide, settings and remove controls appear when you hover it, and a new toggle collapses the whole list.
  • FixedThe AI chat panel's collapse arrow no longer covers the chat history and artifacts buttons when the panel is narrow. Below 400px wide the arrow is hidden; close the chat from the sidebar's AI Chat item or with ⌘I. Long conversation titles now shorten instead of crowding the header.
  • FixedYour plan's monthly AI-credit allowance now resets each billing cycle instead of stacking up, while top-ups and reward credits carry over.
v2.31.0•

Contract roll adjustment

  • ImprovedRoboCharts now adjusts continuous charts for contract rolls, so history no longer jumps at each roll and indicators on the chart use the adjusted prices. A small marker shows each roll. Adjustment is on by default and can be turned off in Chart settings; it switches off while a strategy or order-flow overlay (footprint, order book heatmap, large prints, book levels) is shown. Indicators that load extra data on the server still use unadjusted prices.
  • ImprovedRoboCharts roll markers are now small dots along the bottom of the chart instead of labels on the candles. Hover over a dot, or tap it on touch screens, to see the contract roll and its price gap.
v2.30.1•

Profile picture uploads

  • FixedUploading a profile picture now works.
v2.30.0•

Native strategy template: DTR Range Breakout Return

  • NewDTR Range Breakout Return joins the native strategy templates: after the US open it waits for a candle to close beyond the pre-market range, then trades the close back inside it, shown on an MNQ 1-minute preset with a full backtest you can inspect and clone.
v2.29.1•

AI chat and deployment fixes

  • FixedThe AI chat no longer reports a successful backtest as failed, stops quickly with a clear message when the backtest servers are busy, and no longer shows raw errors for steps it already fixed and retried.
  • FixedDeployments whose bot kept running after a platform interruption again show their strategy, account, starting capital and trade history, and the deployment chart no longer keeps a stale entry line after the bot goes flat.
v2.29.0•

Market matrix optimization

  • NewOptimize runs tick-fill sweeps across several symbols and timeframes at once. With no parameter range enabled and several markets selected, it runs the strategy once per market at its current settings and ranks the markets by your target metric.
v2.28.1•

The AI reads the files you attach

  • FixedFiles you attach to the AI chat now reach the AI's workspace and stay available for the rest of the conversation, so it can open them directly instead of reporting that it cannot read them. This covers trade exports and other CSV files, MQL5 sources and MetaTrader preset files, and common text formats. When an attachment genuinely cannot be loaded, the AI is now told so explicitly rather than assuming the file is there.
v2.28.0•

Strategies with many runnable files

  • NewStrategies and indicators can now hold any number of runnable files. There is no more "main file": Backtest, Optimize, Deploy and the chart ask which file to run (automatically when there is only one), every file compiles on its own, and files can be copied between a strategy and an indicator from the Code tab. The AI never renames your strategies, and a chat upload whose name is already taken is saved alongside the original instead of overwriting it.
v2.27.1•

Live bots roll to the next contract

  • FixedLive bots now pick the correct front-month contract at deploy time across quarterly rolls, pause with a clear message when their contract reaches its roll date, and show the symbol on every order log line.
v2.27.0•

Billing, chat and preset fixes

  • FixedIf your bank declines a plan-upgrade payment, the upgrade no longer stays stuck for a day. You can retry with another card within 30 minutes.
  • FixedUpgrading your plan on a linked account now completes and adds the upgrade credits to your shared wallet.
  • FixedThe Maximize and Fit buttons on the backtest result chart are visible again instead of being hidden behind the price axis.
  • ImprovedThe AI chat toolbar now fits narrow chat panels: attachments, mode, project and model move into a single settings menu instead of being squeezed together.
  • NewFriday Range Revisit joins the native strategy templates: it shorts Monday when the regular-hours open stays inside Friday's range and Thursday's high was never broken, using Thursday's high or an ATR multiple as the stop, with a full backtest you can inspect and clone.
  • FixedStrategy presets handle values outside a parameter's declared range: a saved preset loads exactly as saved and the mismatch shows as a warning instead of blocking the file, and a default that sits outside its own minimum or maximum widens the bound instead of leaving the optimization range without the value the strategy runs with.
  • FixedBilling loads reliably again: it no longer fails or shows the Free plan while your original Roboquant account is being reconciled or the billing service is briefly slow, and background tabs no longer poll billing.
  • FixedIf the renewal payment for your original Roboquant subscription fails, Billing now tells you and lets you update your payment method instead of showing the Free plan.
v2.26.0•

Three new strategy templates

  • NewCharles Monger Investment Strategy by Shujat joins the native strategy templates: a trend-following strategy that goes long when the price flips from below to above a 200-period EMA band, shown on an ES 1-hour long-only preset with a full backtest you can inspect and clone.
  • NewContinuation Havoc joins the native strategy templates: it trades candles that close beyond the previous candle's high or low during a set New York session window, shown on an MNQ 15-minute preset for the first hour after the open, with a full backtest you can inspect and clone.
  • NewThe Double Helix joins the native strategy templates: it enters when price closes beyond a band of high and low EMAs in the direction of a 200-period trend EMA, then trails its stop along the band, shown on an MNQ 1-hour preset with a full backtest you can inspect and clone.
v2.25.6•

AI chat credits and backtest context

  • FixedIf you leave the page, lose connection or press Stop in the first moment after sending a message, the credits reserved for it are now released straight away, so your next message isn't blocked for up to 5 minutes.
  • FixedThe AI no longer assumes a strategy trades ES on the 1-hour chart when you ask it to run a backtest. It uses the market and timeframe you set, the strategy's own market, or asks you.
v2.25.5•

AI chat reliability

  • FixedWhen the AI reaches its step limit on a long task, it now ends with a short summary of what it did, what is left and how to continue, instead of stopping without an answer.
  • FixedPlan mode in the AI chat now works as described: the AI discusses and proposes a plan without editing, saving or running anything until you switch back to Act.
  • FixedThe AI no longer loses a freshly compiled strategy when one of its edits fails or changes nothing, so it can save without recompiling.
  • FixedAttaching a compiled or binary file (for example a MetaTrader .ex5) to the AI chat now explains that it can't be read and suggests attaching the source file, instead of a generic "Could not save the attached file" error.
  • ImprovedThe AI can now point you to where your own AI provider keys are set up: Settings → API keys (BYOK).
v2.25.4•

AI chat credit holds

  • FixedIf a message is interrupted right after you send it (you press Stop, lose your connection or leave the page before the AI starts answering), your credits are released straight away. Your next message is no longer blocked for up to 5 minutes.
v2.25.3•

AI chat credit holds

  • FixedStopping an AI reply, losing your connection or leaving the page mid-reply no longer locks the chat for up to 30 minutes with a "credits are reserved by another AI request" error. Credits held for an AI call that was cut off before it started responding are now released within about a minute.
v2.25.2•

MCP server address

  • ImprovedThe MCP server has moved to https://mcp.roboquant.dev/mcp. Add the server again using the address shown in Connect MCP and approve access once more; your strategies, backtests and data are unchanged.
v2.25.1•

Backtest and billing reliability

  • FixedA plan purchased on Roboquant 2.0 by a customer whose account is linked to a Roboquant 1.0 subscription no longer gets replaced by the 1.0 plan when both are updated at the same time.
  • FixedAI backtests recover saved results when a completion event is lost. Temporary billing checkpoint failures are retried, and credit messages explain when usage settlement is pending.
v2.25.0•

Template discussion author names

  • FixedTemplate discussions now show each commenter's display name instead of "Anonymous".
v2.24.0•

AI chat offers to report problems

  • NewThe AI chat now offers the one-click Report button on its own when a message sounds like something is not working, so you can send a bug report without leaving the conversation.
v2.23.0•

One-click chat bug reports

  • FixedBacktest, strategy and replay charts stay responsive when a strategy draws thousands of shapes: only the shapes in view are drawn while you pan and zoom.
  • NewReport a problem with the AI chat in one click: a new Report button on chat errors and in the composer files a bug report written by Roboquant AI from the recent conversation, no form to fill in.
v2.22.0•

Connector affiliate management

  • NewManage webhook connector affiliates, commissions, and payouts from the Beta admin affiliate panel.
v2.21.2•

CME futures history consistency

  • FixedCME FX and SOFR futures history follows volume-based contract rolls, correcting gaps around quarterly expiries and keeping chart timeframes consistent.
v2.21.1•

Futures rollover and chart-history reliability

  • FixedRecent chart history and backtest results no longer disappear from the platform under load. An internal job queue was growing without limit and was pushing live market-data history out of memory; it is now capped.
  • FixedLive data, charts and newly recorded history for ES, NQ, RTY, MES, MNQ, M2K and the CME crypto contracts now move to the next contract when trading volume shifts to it, instead of staying on the expiring contract until expiry day. This fixes the live chart going quiet on quarterly expiry days. Older stored history around past expiries is being repaired separately.
v2.21.0•

4H FU Retest template

  • New4H FU Retest joins native templates with an MES 15-minute retest preset and a full backtest to inspect and clone.
v2.20.1•

Affiliate experience

  • ImprovedThe affiliate program and dashboard match the homepage design, with clearer coupon sharing, commission balances, referral activity, and payout settings.
v2.20.0•

Affiliate earnings and platform reliability

  • FixedAsking the AI to turn a Pine Script into a Roboquant indicator now saves an indicator instead of a strategy.
  • ImprovedMCP connections support newer clients while preserving existing tools and permissions.
  • FixedConfirming a card wallet top-up no longer fails, and the plan change screen shows the price you currently pay for an original Stripe subscription.
  • FixedRoboquant AI can now run order-book (L2) backtests: strategies that read the order book get real depth data instead of nothing. Longer order-book runs no longer fail partway through.
  • ImprovedFootprint charts build long histories faster, and the Strategies list loads faster for strategies with large files or many versions.
  • FixedSorting optimization runs by a metric through RoboQuant MCP works again instead of returning an error.
  • ImprovedHover or tab onto any backtest, optimization or template metric to see what it means and the exact formula Roboquant uses. Sharpe, Sortino and Calmar are now annualized from each backtest's real time span, and Sortino uses the standard downside deviation, so intraday ratios and Sortino values will differ from earlier runs. Backtest, optimization and template results stored before this change keep their earlier values until they are re-run.
  • ImprovedGrid optimizations with many parameter combinations now keep memory bounded by the top results instead of holding every tested combination.
  • FixedIndicators and strategies no longer get stuck when the wrong file is set as the main file. Compile errors now point at your own files and tell you exactly what to fix, instead of showing a generic server error.
  • FixedAffiliate earnings include eligible Whop subscription purchases and renewals, with refund adjustments and dispute holds. Referral checkout creates a missing personal discount coupon automatically.
v2.19.3•

Free trading tools are back

  • FixedThe free Market Making, Risk Calculator and Monte Carlo tools load again.
v2.19.2•

See your best optimization results at a glance

  • ImprovedThe Optimization Runs list now shows each run's symbol, timeframe and best result (target metric, return, drawdown and trades), and you can rename optimizations directly from the list.
v2.19.1•

Change your original subscription from Billing

  • ImprovedCustomers with an original Stripe subscription can now upgrade, downgrade, or cancel their plan and update their payment method directly from Billing.
v2.19.0•

Roboquant Connect included from Starter

  • NewStarter, Pro, Elite and Ultra plans purchased on Roboquant 2.0 now include Roboquant Connect. Sign in at connect.roboquant.dev with the same verified email to use TradingView webhook execution.
v2.18.2•

MCP connections and strategy reliability

  • FixedCursor can register its native MCP callback, and rejected MCP sign-ins return readable authentication errors.
  • FixedMCP backtests and optimizations use the same default leverage as the app, preventing unexpected zero-trade results from insufficient buying power. Optimization confirmations also show capital, leverage, and execution costs.
  • FixedMCP strategy updates keep Python parameter helpers editable and no longer recompile when only the name, description, or tags change.
  • FixedRoboquant AI accepts image attachments and can take chart screenshots; these options are available again for the default model.
  • FixedA strategy can be redeployed after its bot fails at startup without being blocked as already deployed. Compiled strategies with missing or invalid artifacts show a clear deployment error.
  • FixedVolume profiles retain saved styling, switch periods without briefly adding an extra profile, and show each selected period's color in the layer list.
  • FixedCustom indicators that draw many lines and markers keep rendering on the newest candles of a long chart as older bars are loaded.
  • FixedNew custom indicators start with an empty file tree, ready for your own code.
  • FixedOrder-flow indicators keep their retry controls when another indicator has missing data, and historical order-flow data checks month-to-month continuity more reliably.
  • ImprovedChoosing a paid plan on the pricing page takes you straight to checkout after you sign in or create your account.
v2.18.1•

Free credits and faster first replies

  • ImprovedChat now answers a request for TradingView, MetaTrader/MT5, forex, gold or crypto spot with one sentence on what Roboquant runs plus the closest CME market, instead of attempting a build we cannot run.
  • ImprovedThe first reply in a new conversation is bounded to a fixed number of tool steps, so a single question can no longer run for many minutes before producing an answer.
  • FixedNew Free accounts receive their starter AI credit as soon as they verify their email, so the first message works without opening the billing page first.
  • ImprovedThe upgrade prompt no longer opens the first time you visit Strategies, Templates or Indicators. It appears after your first completed backtest, on that result.
v2.18.0•

RoboQuant strategies on Free, previous account carry-over

  • ImprovedEvery RoboQuant strategy template is now included in the Free plan: the full strategy source is readable and usable without upgrading, and the plan comparison shows it under Free.
  • FixedAccounts created on the previous Roboquant that never verified their email now carry their subscription and wallet balance into Roboquant 2.0 instead of landing on Free.
v2.17.1•

Faster typing in the AI chat

  • FixedTyping in the AI chat no longer lags in long conversations — the composer now updates independently of the transcript.
v2.17.0•

Native strategy template: Previous Day Extreme

  • NewPrevious Day Extreme joins the native strategy templates: it marks the prior CME session's high, low and outer range bands, then trades a close beyond the previous bar's wick out of those extremes with ATR or prior-range stops and targets, with a full backtest you can inspect and clone.
v2.16.2•

Webhooks open Roboquant Connect

  • ImprovedThe Webhooks entry now opens Roboquant Connect at connect.roboquant.dev; the outdated embedded section is gone.
v2.16.1•

Webhook Connector is back

  • ImprovedThe Webhook Connector section is back in the dashboard, sidebar and onboarding for everyone.
v2.16.0•

Roboquant 2.0 is live, faster history loading

  • ImprovedCME history loads faster for charts, custom indicators, and backtests.
  • ImprovedSecond-based chart and backtest history loads faster when bars need to be prepared from archived trades.
  • ImprovedRoboquant 2.0 is live: the beta announcement bar and all closed-beta, waitlist and pre-order wording are gone from the site, sign-in and emails.
v2.15.2•

Deployment stream and workspace start fixes

  • FixedA live deployment's log says why its live stream was interrupted, and a workspace that fails to start reports the real reason right away instead of waiting for a timeout.
v2.15.1•

Chart timeframe switching

  • FixedSwitching chart timeframes keeps candles in view and resets manual price-axis zoom without turning chart padding into large empty areas.
v2.15.0•

Faster chart history

  • ImprovedCharts prepare and deliver candle history faster while preserving the same prices, precision and available history.
v2.14.1•

Large backtest responsiveness

  • FixedLarge backtests with tens of thousands of trades now load quickly and stay responsive on the template page and the Strategies backtest views.
v2.14.0•

Legacy workspace files

  • NewIf you had workspaces in Roboquant 1.0, a new Workspaces section shows them in 2.0 when you sign in with the same verified email. Browse each workspace read-only, download it as a zip, or import its files straight into a Roboquant strategy. Workspaces themselves are retired in 2.0.
v2.13.2•

Chart price scale reset on symbol switch

  • FixedRoboCharts resets the price axis when you switch symbol or timeframe. Previously, after zooming or dragging the price scale by hand, the next symbol kept the old symbol's price range and its candles rendered off-screen.
v2.13.1•

AI answers about strategy engine features

  • FixedThe AI assistant now checks the strategy engine reference before answering whether a feature is supported. It no longer tells you indicators are locked to the close price: moving averages and other single-price indicators accept open, high, low, or combined OHLC sources.
v2.13.0•

Native strategy templates: The Flash Trigger ORB and MNQ Range Breakout

  • NewMNQ Range Breakout joins the native strategy templates: it tracks one targeted 15-minute window a day, then trades the first close beyond its high or low with a Risk:Reward target, a stop-distance filter and an optional reverse trade after a stop-out, with a full backtest you can inspect and clone.
  • NewThe Flash Trigger ORB joins the native strategy templates: it marks the high and low of a single targeted candle and trades the confirmed break with a Risk:Reward target and an optional trailing stop, with a full backtest you can inspect and clone.
v2.12.2•

Replay reliability and heatmap history fixes

  • FixedWide order-book heatmap views load faster and reuse historical data when refreshing, without interrupting unfinished requests.
  • FixedOrder-book heatmaps retain captured intraday history between daily archives and the live feed, including at one-second resolution.
  • FixedOrder-book heatmaps retain recent history longer when live sampling skips seconds, while preserving data gaps and accurate liquidity averages.
  • FixedOrder-book heatmaps show the latest 180 days in wide chart views instead of a date-range error.
  • FixedOrder-book heatmaps automatically adjust price grouping to fit wide views with fragmented recent history.
  • FixedReplay mode shows every candle again on long runs; the result stream no longer drops replay frames under load.
v2.12.1•

Workspaces stay current after updates

  • FixedWorkspaces run the latest platform version right after each update instead of an older cached one.
v2.12.0•

Native strategy templates: Crazy ORB and Williams %R Long

  • NewTwo new native strategy templates are available in Templates: Crazy ORB, a close-confirmed New York opening-range breakout, and Williams %R Long, a long-only Williams %R mean-reversion model, each with a full backtest you can inspect and clone.
v2.11.0•

Chart controls, 2m timeframe and replay bars

  • ImprovedMobile charts have scrollable menus, visible replay controls, full-width Book and Tape views, searchable drawing tools, and clearer statistics and settings layouts. Tape resumes at the newest trade after scrolling.
  • NewThe 2-minute timeframe is available for charts, backtests, optimization and bots.
  • FixedThe AI chat can create, save and backtest a strategy while an indicator is bound in the session.
  • FixedIndicator markers no longer print the indicator name next to every dot; the text is now opt-in with axis_label_visible=True (Python and compiled Rust indicators).
  • FixedBacktest replay mode animates the candles again instead of jumping straight to the results.
v2.10.0•

Campaign pages, order-book heatmap and open signup

  • ImprovedTimestamps that show both a date and a time use a comma in every browser, for example "15 Jan 2026, 15:30" instead of "15 Jan 2026 at 15:30".
  • NewUltra users can add an order-book heatmap to RoboCharts for supported continuous CME instruments, with zoom-dependent history down to one-second detail where available, time-weighted averages when zoomed out, live resting liquidity, bright liquidity concentrations, smoother navigation, bid/ask hover values, replay, and adjustable grouping and colours.
  • FixedScrolling over the right-hand price scale in RoboCharts adjusts that pane's vertical scale without zooming or panning the time axis.
  • NewExplore focused introductions to building and testing strategies, CME data, cloud optimization and the free RoboQuant 2.0 course. Each page explains the free plan and relevant paid limits.
  • NewBacktest reports show hourly net P&L after commission, across all days or for a selected day, using the report timezone.
  • FixedWhen public signup is open, the beta-access page sends you directly to create a free account. The old /auth link now opens the sign-in and signup page.
v2.9.0•

Mobile layouts, faster chat and dictation

  • NewIn Beta AI chat, dictate messages with live text in the draft, choose your microphone, and test its input locally. Dictation uses AI credits; review your message before sending.
  • ImprovedMobile layouts use compact, consistently aligned controls and keep content clear of the bottom navigation. RoboCharts has a translucent floating toolbar, with trading, strategy results, and replay available from its menu. Strategy configuration, indicator cards, billing, and documentation also fit smaller screens.
  • FixedTyping stays responsive in long AI conversations. Backtest results and charts open on demand so saved chats load with less data.
v2.8.0•

Individual CME expiries and shared symbol search

  • NewBrowse individual CME futures expiries alongside continuous markets in RoboCharts, Backtest and Optimize. Available individual-contract history supports OHLCV charts and simulations from one minute upwards, and saved runs retain the contract history they used. Continuous symbols are still required for sub-minute timeframes, tick and order-book fills, and live deployments. Footprint, Delta and CVD also remain available only on continuous symbols.
  • ImprovedFind CME futures in RoboCharts, backtests, and optimizations with compact symbol search, distinct market icons, an expanded market browser, and shared favorites.
v2.7.0•

Clearer charts, compact controls and reliable chat

  • ImprovedThe platform has more compact typography, controls, and spacing, with a streamlined dashboard, wider template cards, and direct access to models and API keys in Settings.
  • ImprovedCME footprint charts on eligible plans show buy and sell volume beside filled candles, with larger numbers and an outline around each candle's highest-volume row. Hover over a price row for exact bid, ask, volume, delta, and tick-range details; a bottom table shows whole-candle totals. Ticks-per-row settings stay fixed while zooming, and delta consistently shows ask volume minus bid volume.
  • FixedChat reports interrupted responses and restores your message for retry, while rapid repeat sends no longer start duplicate requests. OpenAI reasoning controls include Astra Max. On plans with support for your own AI keys, OpenAI chats preserve images and reasoning during tool calls and report streaming usage correctly.
  • FixedVerified Free accounts linked to an existing RoboQuant identity receive their one-time starter credits when their shared wallet has no balance, deposits, or spending history. Claude chats with Thinking on avoid unnecessary full-context credit reservations, and GLM models work again. All models use a 10-cent initial wallet check; each request still needs enough credits for its estimated cost.
  • FixedChoose which Tradovate sub-accounts to enable within your plan allowance. Free includes no connections or native deployments; Starter allows one Demo connection and no native deployments. Pro, Elite, and Ultra allow 2, 5, and 20 enabled accounts and 1, 3, and 7 native deployments respectively, with Demo and Live sharing each allowance. Existing enabled accounts and running bots remain usable after a downgrade. Billing, Support, and deployment docs show consistent allowances, and billing and checkout receipts no longer show legacy workspace limits.
  • FixedThe Optimize market picker scrolls through every symbol and timeframe, fits smaller screens, and has a clear screen reader label.
  • ImprovedThe Webhook Connector section is hidden on Beta, and saved links to it open the dashboard.
v2.6.0•

Faster swing and pivot strategies

  • FixedDashboard pages scroll normally on phones and tablets, including when swiping over tables.
  • ImprovedSwing and pivot strategies that rescanned their whole history on every bar could take hours to backtest and days to optimize. The strategy SDK now offers a bounded lookback window (Rolling), the AI writes per-bar-efficient strategy code, and a new parity harness guarantees such rewrites produce identical trades. A typical 9000-combination optimization on a year of 5-minute bars now finishes in under two minutes.
v2.5.0•

Heikin Ashi and custom candle colors

  • NewCustom indicators can plot Heikin Ashi and other candle series over the price chart or in a separate pane, and highlight individual market candles by condition.
v2.4.2•

Content rewards and indicator fixes

  • ImprovedFind Content rewards through the Earn a free subscription link in your profile menu. View milestones for 50% and 100% refunds now depend on your plan and require one video on one platform; views from other videos, platforms, shares, or reposts cannot be combined.
  • FixedNew indicators are available in RoboCharts immediately. AI edits stay linked to the saved indicator, and compile retries retain its name.
  • FixedAI indicator previews report band data correctly and use the custom indicator API reference when writing code.
v2.4.1•

Strategy deployment and indicator fixes

  • FixedDeployment now uses the selected backtest’s market, timeframe, parameters, and entry file instead of silently using the latest run.
  • FixedUpload files from the indicator code tab, remove unwanted text annotations, and get accurate point counts when the AI previews an indicator.
v2.4.0•

Learning path and clearer model selection

  • NewLearn RoboQuant with 13 free video lessons inside the app. Resume lessons across devices, follow your first strategy and backtest milestones, and open walkthroughs from Strategies and Accounts. The full free course is also available on Whop.
  • ImprovedThe chat model picker has clearer model details and selection, a compact reasoning selector, and an easy-to-find model library button.
v2.3.0•

Discord community and upgrade checkout

  • NewJoin the Roboquant Discord community straight from the sidebar, on desktop and in the mobile menu.
  • ImprovedUpgrading your plan now shows a checkout summary before anything is charged: the prorated amount, the credit for the unused part of your current plan based on the rate you actually pay, the card on file, the next renewal date, and a promo code field.
  • FixedAI chat replies are no longer cut off by an interrupted turn or a Roboquant update. Answers in progress keep streaming while a new version rolls out, interrupted requests release their credit hold automatically so chat no longer gets stuck on "Insufficient reserved funds", each request reserves far fewer credits up front, and credit errors now explain what is wrong and how to fix it.
  • FixedOpening a conversation with an older, incompatible saved tool message no longer closes the dashboard; the saved content stays available to inspect.
v2.2.0•

Your AI model library

  • NewChoose from 30 AI models and save your chat shortlist to your account, including Claude Fable 5.1, Muse Spark 1.3, DeepSeek V4.1 Flash, and MiMo V2.5 Pro.
  • ImprovedFind models in a compact library grouped by company. Switch models on instantly and check capabilities and estimated credit costs from the info icons.
  • FixedModel logos display in both themes, and leaving the library while preferences load no longer raises a cancellation error.
  • NewEligible paid plans can upgrade immediately with a prorated charge or schedule a downgrade for the end of the billing period. Cancel a scheduled change from Billing.
  • FixedPlan-restricted backtesting, optimization, and deployment actions show upgrade options instead of generic errors.
  • FixedLong bug descriptions scroll within the Report a Bug dialog, keeping its title and Submit button reachable.
v2.1.12•

Live deployment startup fix

  • FixedLive deployments started since Sep 9 stopped within minutes with no reason. The trading runtime image was missing its live-session engine; it is restored, the image build now refuses to ship without it, and a deployment that cannot start now shows Error with the cause instead of a silent Stopped.
v2.1.11•

Backtest handoff and RoboCharts navigation

  • FixedOptimization results open a fresh backtest with saved settings and the selected market. New results preserve the evaluated bar window, with separate in-sample and out-of-sample reruns. Reopened backtests retain their saved costs and entry file. Backtests no longer include the next day's opening candle.
  • FixedBacktest metrics adapt to the available panel width, keeping numbers readable in AI chat and when resizing the chat beside a report.
  • FixedRoboCharts keeps indicator panes aligned when dragging or zooming and preserves the visible period across timeframe changes and indicator loading.
  • ImprovedRoboCharts shows a labeled recent-history strategy preview while longer OHLCV backtests finish, with clearer loading stages and a live elapsed timer instead of misleading completion estimates.
v2.1.10•

Chat and chart reliability

  • FixedChats no longer reject requests solely because encoded attachments exceed a token limit in bytes, and image attachments no longer trigger unnecessary history compaction based on their encoded size.
  • ImprovedStrategies can choose open, high, low, close, or combined OHLC prices for moving averages, momentum and statistical indicators, MACD, and Bollinger Bands.
  • FixedRoboCharts keeps sparse indicators aligned with candles. Long and short position drawings start at 1:1 risk/reward with two clicks and update smoothly while dragging their entry, target, or stop.
  • FixedRoboCharts avoids redundant indicator calculations when switching timeframes and keeps results aligned with the selected candles.
  • FixedLive updates such as backtest progress, file sync, strategy locks, and deployment status keep arriving after your session token refreshes, instead of silently stopping in tabs left open for more than an hour.
  • FixedAffiliate links stay on the site where the affiliate account was created, so Beta referrals no longer send buyers to the production site.
  • ImprovedAffiliates can choose a personal coupon that gives new customers 10% off their first subscription payment. Referral links apply the coupon automatically, and entering it at checkout records the referral.
v2.1.9•

AI request availability

  • FixedInterrupted AI requests reserve their pending costs while leaving remaining credits available for new requests. DeepSeek checks account for the submitted prompt, and long responses keep the connection responsive.
v2.1.8•

Prompt-sized credit checks

  • FixedClaude text and tool requests with thinking off check the size of your prompt, so smaller credit balances can start affordable requests.
v2.1.7•

Delayed AI usage recovery

  • FixedInterrupted AI requests retry delayed usage receipts automatically, releasing the reservation after confirmed usage is settled.
v2.1.6•

Interrupted request credit recovery

  • FixedInterrupted AI requests check confirmed provider usage before settling credits, recovering reservations when the final usage response was missed.
v2.1.5•

Long conversation credit reliability

  • FixedLong AI conversations correctly account for context summaries and completed retry attempts, releasing their unused credit reservations.
v2.1.4•

AI credit availability

  • FixedAI requests with multiple tool steps reuse unused credit reservations after each confirmed response, preventing false insufficient-funds errors on funded accounts.
v2.1.3•

AI request recovery

  • FixedInterrupted AI requests recover confirmed usage without leaving credits locked, and reopened tool histories no longer fail because of missing tool-call links.
v2.1.2•

Production plans in Beta

  • FixedYour Production plan and AI credits carry over to Beta when both accounts use the same verified email. Both apps use one shared credit balance.
  • ImprovedContent rewards now uses the updated RoboQuant design, with a more accessible submission form, searchable review queue and clear decision confirmations.
v2.1.1•

Wallet payment validation

  • FixedWallet top-ups are credited only after the paid amount and tax values are validated.
v2.1.0•

Content Reward Program

  • NewSubmit a video through the Content Reward Program for a manually reviewed refund: 50% at 2,000 views or 100% at 5,000 views.
v2.0.2•

Native template reliability

  • FixedRoboQuant templates compile automatically when you select Use this template, then open Backtest with the compiled strategy selected.
v2.0.1•

AI wallet billing reliability

  • FixedAI wallet billing preserves fractional usage charges and prevents duplicate payment credits. Paid chats reserve funds before starting; requests with unconfirmed usage keep those funds reserved for reconciliation.
  • NewNew customers following an approved affiliate link receive 10% off their first Starter, Pro, Elite, or Ultra subscription payment. Renewals use the regular price.
v2.0.0•

RoboQuant 2.0 — Build, Backtest & Trade

  • NewRoboQuant 2.0 brings strategy creation, native backtesting, optimization, charts, and deployment into one redesigned dashboard
  • NewNative RoboQuant strategies — create and compile your strategy, then manage its Backtest, Optimize, History, and Deploy workflow from Strategies
  • NewNative backtesting runs directly on Databento/CME futures data, with bar-based, tick-level, and order-book fills available according to your plan
  • NewBacktest multiple symbols and use sub-minute timeframes, with configurable commissions, slippage, position sizing, and trailing stops
  • NewBacktests and optimizations keep running in the background — reopen a run to follow its progress, live equity curve, and logs, or cancel it
  • NewExpanded backtest reports include trade-by-trade results, cumulative P&L, monthly returns with yearly totals, and performance by period and weekday
  • NewDrawdown analysis shows each decline, its duration, and recovery; risk reports add Monte Carlo simulations, confidence intervals, and trade-excursion analysis
  • ImprovedProp firm analysis includes richer challenge presets, risk controls, and evaluation of strategy results against challenge rules
  • NewNative optimization offers grid, Bayesian, genetic, and multi-objective searches, plus out-of-sample and walk-forward validation on eligible plans
  • NewOptimize across multiple markets, inspect the full results table and 3D parameter surface, then apply selected settings to a new backtest
  • NewSave and reuse parameter configurations across backtesting, optimization, and deployment, with portable settings files
  • NewRoboCharts brings interactive CME futures charts, saved layouts, built-in indicators, custom colors, and configurable chart panels to the dashboard
  • NewOrder-flow tools add footprint charts, volume profiles, cumulative delta, large trades, absorption, and live order-book and trade-tape views on eligible plans
  • NewChart replay supports bar-by-bar and tick-by-tick playback, with entry, take-profit, stop-loss, and trailing-stop levels visible as trades unfold
  • NewAttach a strategy to RoboCharts to inspect its trades, indicators, and drawings, then backtest, optimize, or deploy from the chart
  • NewCreate and edit custom indicators, manage their versions, and add them to charts alongside built-in technical indicators
  • ImprovedChart tools include more drawing shapes and line styles, indicator visibility toggles, multiple volume-profile periods, and editable color codes
  • NewNative strategy templates include backtest results so you can inspect performance before creating your own strategy from a template
  • ImprovedStrategy editing supports multiple files and folders, inline AI changes, version history and restore, and searchable lists with pagination
  • NewDeploy strategies to connected Tradovate demo or live accounts on eligible plans, with configurable inputs and risk limits
  • NewMonitor deployments with live charts, positions, orders, trade history, and strategy P&L; pause, resume, update parameters, or stop a deployment
  • ImprovedThe trading panel groups positions, orders, and trade history by account, with controls for managing orders and positions
  • ImprovedThe AI chat dock follows your selected strategy or indicator, shows clearer activity and file changes, and lets you queue messages while the AI works
  • ImprovedAI chat supports PDF and CSV attachments, turns large pasted text into attachments, and displays interactive charts directly in the conversation
  • NewConnect your own supported AI provider keys on eligible plans, with key validation, pause controls, and separate usage reporting
  • NewConnect external AI assistants through RoboQuant MCP on eligible plans to work with native templates, strategies, backtests, optimizations, and deployments
  • ImprovedBeta billing brings the Free, Starter, Pro, Elite, and Ultra plans together with AI credit top-ups, saved payment methods, auto-recharge, and in-app cancellation
  • NewVerified Free members receive 2,000 starter AI credits once; paid plan credit allowances continue to renew monthly
  • NewThe affiliate dashboard lets enrolled members track referrals, commissions, and payouts
  • ImprovedDashboard navigation, charts, forms, and result tables work better on mobile, with touch controls, clearer notifications, and Help and Report a Bug actions in the header
  • ImprovedSign-in guides you through email verification and required CME subscriber forms, while charts show feed status and delay labels
  • FixedReopened backtests preserve their fill method, settings, charts, and trade markers; finished backtests and optimizations can now be deleted from history
  • FixedLong backtests deliver results faster, optimization progress remains visible across validation stages, and failed runs show errors instead of staying stuck as running
  • FixedChart replay starts more reliably, footprint history stays continuous, and indicator settings update without unnecessary reloads
  • FixedAI chat recovers more reliably from interrupted connections, preserves attachments, and opens the correct prompt from New Strategy and Ask AI actions
  • FixedStrategy and indicator saves stay in sync, AI edits run against the current code, and file operations no longer create duplicate strategy items
  • FixedDeployment trade history and P&L stay scoped to the correct strategy and account, with more reliable broker reconnection and account selection
  • FixedBilling handles declined auto-recharges correctly, preserves purchased plan access, and labels Free starter credits as one-time rather than monthly
v1.23.1•

Smarter AI & Improved Reliability

  • ImprovedRoboQuant AI is now smarter and faster — improved coding quality with better long-task reasoning
  • ImprovedAI now automatically retries on a backup engine if the primary is busy — no more random disconnections
  • FixedAI chat no longer stops mid-response with "network error" during high traffic
  • FixedStrategy Optimizer correctly resets parameter selections when reopening
v1.23.0•

RoboQuant AI — Now with Vision

  • NewRoboQuant AI now supports image uploads — attach screenshots, charts, or documents and get AI analysis directly in chat
  • NewExtended thinking mode for RoboQuant AI — deeper reasoning for complex strategy logic
  • ImprovedRoboQuant AI is significantly faster — upgraded to a new turbo engine for snappier responses
  • ImprovedRoboQuant AI can now generate longer responses (up to 131K tokens)
  • FixedElite plan users can now access the Webhook Connector and TradingView Optimizer without being blocked
  • FixedWebhook broker account selection now saves correctly after refresh
  • FixedTrading calendar no longer shows phantom trades with incorrect P&L
  • FixedBracket orders (SL/TP) now use correct tick-based price precision
  • FixedCancel all orders now works reliably on Tradovate
v1.22.0•

Prop Firm Challenge Simulator

  • NewProp Firm Challenge Simulator — test your optimization parameters against real prop firm rules (Apex, TopStep, MFF, FTMO, Trading Pit) and see your pass probability, net expected value, and optimal risk multiplier
v1.21.0•

Webhook & Trading Calendar Fixes

  • FixedTrading calendar shows accurate P&L — entry prices now correctly match the traded symbol
  • FixedClose trades are properly recorded in your trading calendar
  • FixedPosition matching works correctly when trading multiple symbols simultaneously
  • FixedSwing trades held for multiple days now appear correctly in the calendar
v1.20.0•

Webhook Per-Account Sizing & Notifications

  • NewWebhook connector now supports per-account quantity multipliers — set different position sizes for each trading account in one alert
  • NewEmail notifications for webhook order fills and errors — get notified when your trades execute or fail
  • NewProp firm challenge analytics include risk multiplier and auto-normalize P&L to your account size
  • ImprovedWebhook connector test button now works for all broker types (Tradovate and crypto exchanges)
  • ImprovedWorkspace sessions stay active longer (up to 16 hours) with better hibernation warnings
  • FixedWebhook delivery now works with all three alert payload formats — no more "400 Bad Request" errors
  • FixedBroker connections stay active more reliably — automatic token refresh handles edge cases better
  • FixedWorkspace billing no longer loops on authentication errors
v1.19.0•

Artifact Version History

  • NewBrowse all versions of AI-generated files in the sidebar — expand any file to see its full edit history with timestamps
  • NewName your file versions — click any version label to add a custom name (e.g., "Added RSI filter")
  • ImprovedAI chat messages show one card per file instead of duplicates when the same file is edited multiple times
  • ImprovedTradingView extension supports manual JSON import for optimization results
  • FixedAI file editing works correctly with files that contain Windows-style line endings
  • FixedSeveral minor production errors resolved
v1.18.1•

Stability Fixes

  • FixedAI chat no longer loses context during long conversations with many tool calls
  • FixedFile saves are more reliable under high concurrency
v1.18.0•

AI Chat Reliability & Extension Upgrade

  • ImprovedAI chat is more reliable — no more "Network Error" during long code generation sessions
  • ImprovedBilling accuracy improved — usage is now correctly tracked even if you close the tab mid-conversation
  • FixedFile preview and diff editor now work correctly in dashboard chat
  • FixedTodo list UI restored in dashboard chat
  • FixedWebhook connector: Tradovate token refresh is faster and more reliable — fewer disconnections
  • ImprovedTradingView extension handles large optimization results without crashing (v2.1.3)
  • ImprovedTradingView extension now supports dropdown and boolean strategy inputs
  • FixedPassword reset flow works correctly again
  • FixedAI companion input text no longer overflows its container
v1.17.1•

Claude Sonnet 4.6

  • NewClaude Sonnet 4.6 is now the default AI model — faster responses with improved intelligence
v1.17.0•

TradingView Strategy Optimizer

  • NewTradingView Strategy Optimizer — optimize your strategy parameters directly from TradingView with real-time results
  • NewOptimization history — view and compare all your optimization runs from Strategies
  • FixedCSV backtest upload now works with non-USD currencies (CHF, EUR, GBP, and more)
v1.16.0•

Prop Firm Challenge Simulator

  • NewProp firm challenge simulator: see your probability of passing evaluation rules
  • ImprovedAI chat error messages are clearer and more helpful
v1.15.0•

MT5 Backtest Import & AI Companion

  • NewMetaTrader 5 backtest report import (legacy feature, since retired)
  • NewAI Companion widget — ask questions about your data from any page
  • FixedMonthly free credits now renew reliably for all members
  • FixedCSV backtest upload errors resolved
v1.14.0•

Claude Opus 4.6

  • NewClaude Opus 4.6 is now available — Anthropic's most intelligent model with 128K output
  • ImprovedOpus pricing reduced by 66% — now 3x cheaper than before
v1.13.1•

MT5 Compiler Fix

  • FixedAI can now compile MetaTrader 5 code reliably
  • ImprovedAI chat retries automatically up to 5 times if interrupted
v1.13.0•

Webhook Bracket Orders

  • NewBracket orders with Take Profit and Stop Loss - set TP/SL in your TradingView alerts
  • NewTick-based or price-based brackets - specify offsets in ticks or absolute prices
v1.12.1•

AI Chat Reliability

  • ImprovedAI chat preserves your conversation when switching models or getting interrupted
  • ImprovedContext usage display is more accurate and updates in real-time
  • ImprovedLarge files are handled smarter - AI uses partial reads to avoid memory issues
  • FixedAI works better with Chinese AI models (GLM-4 formatting issues)
  • Fixed"Report a Bug" button restored in header
v1.12.0•

AI Chat V2 & Backtest Analytics

  • NewAI Chat V2 - Redesigned chat with instant startup and file attachments
  • NewBacktest analytics with Monte Carlo simulations and detailed performance metrics
  • NewStrategy Templates - Browse and use pre-built trading strategies as starting points
v1.11.1•

Webhook Safety & Reliability

  • ImprovedWebhook orders now require account selection - prevents orders going to unintended accounts
  • ImprovedTrade protection blocks closing orders if entry failed - prevents accidental naked positions
  • ImprovedMulti-account webhooks are more reliable - fixed rate limiting when trading same symbol
  • FixedTrade Calendar now shows all webhook trades with accurate P&L
v1.11.0•

Crypto Trading & Trade Calendar

  • NewTrade Calendar - visualize your trading P&L by day, week, or month with detailed fill history
  • NewKraken exchange support - connect Kraken Spot and Kraken Futures accounts
  • NewBinance, Bybit, and Hyperliquid crypto exchanges now available
  • NewReal-time position updates - see balance and P&L changes instantly via WebSocket
  • ImprovedWebhook execution is now ~4x faster
  • ImprovedWorkspace stays connected when switching dashboard sections
  • ImprovedDashboard loads faster - positions refresh in background
v1.10.0•

Codex 5.2 - Frontier Coding Model

  • NewCodex 5.2 is now available - the frontier model for long-running coding tasks
v1.9.1•

Bug Reporting

  • NewReport a Bug button - click the bug icon in the header to send feedback directly to our team
v1.9.0•

Webhook Connector - Auto-Trade from TradingView

  • NewWebhook Connector - connect TradingView alerts to Tradovate and auto-execute trades
  • NewMulti-account trading - send orders to multiple Tradovate accounts simultaneously
  • NewRisk management controls - set position limits, daily order caps, and emergency kill switch
  • NewSymbol whitelist - restrict trading to specific futures contracts only
  • NewOrder logs - track every executed trade with full details and status
  • NewChart Analyzer history - your analysis conversations are now saved and accessible
  • ImprovedSmart symbol detection - TradingView continuous contracts (ES1!, MNQ1!) now work automatically
  • FixedLarge Tradovate order IDs now track correctly in order history
v1.8.0•

TradingView Extension & AI Streaming

  • NewTradingView Bridge Extension - validate Pine Script directly in TradingView with real compiler results
  • NewOne-click validation - click "Validate" and see real TradingView errors instantly in the chat
  • New"Fix with AI" button - when validation finds errors, let AI fix them automatically
  • NewExtension auto-opens Pine Editor and compiles your code seamlessly in the background
  • ImprovedAI responses now stream smoothly in real-time instead of appearing word-by-word
  • ImprovedSee AI reasoning as it thinks - watch the thought process unfold live
v1.7.0•

GLM-4.7 - New Default AI Model

  • NewGLM-4.7 is now the default AI model - faster responses and better code generation
  • NewModel selector now shows pricing per million tokens - know the cost before you choose
  • ImprovedMore accurate billing - costs now match actual AI usage precisely
v1.6.1•

AI Reliability Improvements

  • FixedAI no longer gets stuck repeating the same step - now detects loops and asks for guidance
  • FixedAI now follows through when it says "Let me do that" instead of stopping early
  • FixedDuplicate messages no longer appear after reconnecting to a chat session
  • ImprovedMore reliable task completion for complex multi-step requests
v1.6.0•

New AI Models & Video Tutorials

  • NewGemini 3 Pro & Gemini 3 Flash models - Google's latest AI with excellent coding capabilities
  • NewGrok 4.1 Fast model - xAI's lightning-fast model for quick iterations
  • NewVideo tutorials - embedded tutorial videos in the dashboard Overview tab
  • NewWelcome tutorial modal - new users get an onboarding video on first visit
  • NewChart Analyzer documentation - learn how to analyze charts with AI
  • ImprovedModel selector now shows provider icons (Anthropic, OpenAI, Google, xAI)
  • ImprovedResources section updated with current platform documentation
  • ImprovedFAQ updated with latest plan pricing and features
  • FixedPreview browser now supports scrolling for web apps
v1.5.0•

New Wallet System & Usage-Based Billing

  • NewWallet system - pay only for what you use with transparent per-request pricing
  • NewReal-time cost tracking - see exactly how much each AI request costs
  • NewAuto-recharge - automatically top up your wallet when balance is low
  • NewPayment methods - save cards for quick deposits and auto-recharge
  • NewUsage dashboard - detailed breakdown of AI usage by model and cost
  • NewMulti-tier Whop support - Premium, Advanced, and Ultra memberships
  • ImprovedAI models updated to Claude 4.5 versions (Haiku, Sonnet, Opus)
  • ImprovedDemo workspace now uses Claude Haiku 4.5 for faster responses
  • FixedBilling accuracy - costs now calculated based on actual token usage
v1.4.0•

OpenRouter Fallback & GLM-4.6 Model

  • NewGLM-4.6 model - a powerful Sonnet-class coding model with 200K context, full tool execution, and excellent code generation at no extra cost
  • NewOpenRouter fallback - Claude models automatically route through OpenRouter when Anthropic rate limits are hit
  • ImprovedMore reliable AI responses during high-traffic periods with automatic provider switching
  • FixedWorkspace connection issues during high traffic - auto-scaling now handles load spikes better
  • FixedTool execution reliability for all supported models
v1.3.0•

Video to Strategy & GPT-5.2 Models

  • NewVideo transcript extraction - paste a YouTube, TikTok, Instagram, Twitter, or Facebook URL and AI generates trading strategies from it
  • NewGPT-5.2 and GPT-5.2 Pro models with advanced reasoning capabilities
  • NewReasoning effort controls for OpenAI models (low, medium, high)
  • NewDirect backend streaming bypasses timeout limits for long AI sessions
  • ImprovedExtended thinking support for complex code generation tasks
  • FixedConnection timeouts during long code generation sessions
v1.2.0•

Multi-Workspace & Session Management

  • NewCreate multiple workspaces with different trading strategies
  • NewSession hibernation saves resources when inactive
  • ImprovedFaster container startup times
  • ImprovedBetter file sync between editor and container
  • FixedFile changes now persist correctly across sessions
v1.1.0•

Enhanced AI Chat & Code Execution

  • NewAI can now execute Python code directly in your workspace
  • NewPine Script syntax highlighting and validation
  • NewKnowledge base search for trading strategies
  • ImprovedMore accurate code generation with context awareness
  • FixedTerminal output now displays correctly

Want to try the latest features?

Start building and automating your trading strategies today.

Get Started