AI Chat

AI Chat is one assistant for everything you do before risking capital: ask about markets, research a question on the web, build or refine a strategy, and backtest it, all in the same conversation. There's no separate "trader" or "builder" mode to pick: just type.

The same AI Chat is in the Live section too (Trader and Quant), in Live's blue colours. It's the same assistant with the same conversations: switching between Testing and Live keeps the chat you have open.

Not financial advice

AI analysis can be incomplete or wrong. Every answer is a starting point for your own judgement, not a recommendation to trade.

Ask about markets and your work

Open AI Chat from the sidebar and ask a question, for example:

What's moving in crypto today?
Review my strategies.
Why did my last backtest lose money in March?

On the Trader and Quant plans the assistant can read your own work (saved strategies, backtest results, paper runs) and current market context, and it shows where each fact came from under the answer. On Free, questions are answered with web research only (below).

When a reply ends with suggested follow-ups, tap one to send it as your next message.

Charts, tables and cards

On Trader and Quant, answers about your work can include cards under the text, for example:

metric tiles for a result (return, max drawdown, Sharpe, trades);
an equity curve with a drawdown chart under it;
a price chart with the trades marked on it;
a Monte Carlo fan chart and a correlation heatmap;
tables of runs, trades or strategies. Select a column header to sort by it, and Copy CSV copies the rows shown;
cards for runs and strategies, with a small equity line;
news and upcoming events (times in UTC). News links open in a new tab;
strategy code, with Copy.

The numbers on cards come from AlphaProve's own results and market data, not from the AI. Backtest figures are marked Hypothetical: they're simulated, not real trading. Open on a card goes to the full page, such as the backtest's tearsheet. Select a card's title to fold it away.

Some cards end with suggested next steps. A button can open a page, open a form already filled in (for example Run a forward walk), put a question in the message box or send a follow-up. A form still waits for you to submit it, so nothing runs from a button. Export to Pine opens the export window and waits for you to select Export before it spends AI credits.

If a card says Unsupported view, reload the page to get the latest version of the app.

Open a page from chat

Ask "Open my last backtest" or "Take me to the Monte Carlo check for this run" and the assistant opens that page in the same tab. It only opens a page by itself when your message asks to open or go somewhere. Otherwise it offers a button, such as Open the result. Reloading a conversation never opens a page again.

It can open your Library and strategies, the editor, Results and tearsheets, forward walks, Optimizer Lab campaigns and reports, paper runs, the Risk page, the Live Terminal and Settings. Links only ever point at your own items. In Live, a page opens in Live when Live has that page.

The copilot dock

You don't have to leave a page to ask about it. Press ⌘J (Mac) or Ctrl+J (Windows and Linux), or select Ask AI at the bottom of the sidebar. The chat opens at the side of the page (at the bottom on a phone).

The header shows Looking at: and the page you're on. On Trader and Quant, the assistant knows which strategy, run or section that is, so "Why did this lose in May?" is enough. It checks that each item is yours before it reads it.
It continues the conversation you last opened. New chat starts another.
Expand moves the conversation to the full AI Chat page.
Runs you start from the dock open on their own results page.
Press Escape in the dock, or ⌘J / Ctrl+J again, to close it. The dock isn't shown on the AI Chat pages themselves.

Press ⌘K (Mac) or Ctrl+K anywhere in the app for the command menu. Ask AI about this page opens the dock. The same menu goes to any page or recent conversation.

Monte Carlo and prop-firm checks

On Trader and Quant, ask about the risk of a finished backtest, for example "Run a Monte Carlo check on my last BTC backtest" or "Which prop firms would this backtest pass?"

Monte Carlo reshuffles the backtest's trades into 2,000 paths. The answer quotes:
the risk score;
the chance of a drawdown deeper than the ruin threshold;
the median worst drawdown;
the worst drawdown in dollars that only 1 path in 20 exceeds.

The dollar figure uses the backtest's starting capital unless you name an account size. Reshuffling keeps the final return fixed, so the chance of profit isn't estimated here. The Risk page's other methods estimate it. - Prop firms replays the backtest against the prop-firm catalog. The answer lists the best 10 programs, each with pass, open or fail and the first rule that failed. It also says how many programs passed, are still open and failed. - Full results, charts and settings are on the Risk page, linked under the answer. Like the Risk page, checks are limited per minute. They use no AI credits beyond the answer itself. - These checks only read results. They never start a run or make a plan.

Research on the web

Ask for news or facts and the assistant can search the public web and read a few pages before it answers.

Cited sources. Facts from the web carry markers like [W1]. Select one to jump to its source card, which shows the page title, the site, when it was published (or "Publication date unknown") and when it was retrieved. Cited sources come first. Pages it looked at but didn't cite are under Also retrieved. Some primary sites (governments, central banks, regulated exchanges, company investor relations) are labelled Official source.
Links you paste. Paste an https:// link and ask about it; the assistant reads that page directly.
Per answer: up to 2 searches and 3 page reads. An answer that uses web evidence costs one extra AI credit. A search that finds nothing costs no credit but still counts toward your monthly searches.
Monthly search allowance (resets on the 1st, UTC):

Plan Web searches per month

Free 30

Trader 150

Quant 500

When 10 or fewer are left, the remaining count appears under the sources. When none are left, the assistant says so and when the allowance resets. Reading links you paste still works. Occasionally search is busy for everyone; try again later.

What it will never do:

Invent a source. If search is unavailable or finds nothing, it says so. A marker that doesn't match a retrieved page is flagged as unverified.
Open private or internal addresses. It only reads public https:// pages from its own search results or from links you pasted.
Send your private data to the search engine. It refuses searches that repeat your strategy names or account figures from the conversation.
Follow instructions written on a web page. Page text is treated as information. Nothing on a page can start, change or accept anything; plans still need your acceptance.

Build and refine a strategy

Describe an idea in plain English (an entry rule, a market, a timeframe) and the assistant asks about anything missing, then writes the strategy when you select Generate code. Keep chatting to change it; a research question in between doesn't disturb your draft. See Getting started for the full walkthrough and Strategy authoring for writing strategies by hand.

Improve a saved strategy

Choose Improve in chat on a strategy in your Library or on a backtest result, or pick it with Strategy in the message box. Then say what you want, for example "Lower the drawdown without changing the entries." The assistant agrees a short brief with you:

Goal: what the change should achieve.
May change: which parts the assistant may touch. Choose any of Parameters, Entry & exit rules, Filters & indicators and Risk & sizing.

Select Generate candidate. The assistant writes a new version, a candidate, and your original strategy stays exactly as it was. A candidate is created only when its code passes validation and stays inside the brief. For example, if Risk & sizing isn't ticked, the candidate's RISK block can't differ from the original's.

The candidate card shows:

the assistant's list of changes and why;
the code changes, compared with the version it came from or with your original;
Run backtest, Add to Library and Apply to original.

Keep chatting to refine it further: each new candidate builds on the last one, and all of them stay listed under the original on its Library page. Regenerate on a candidate's reply makes another candidate from the same starting point; the earlier one stays in your Library. Candidates join the original's family, so the Optimizer Lab counts their trials together.

Apply to original replaces your original's code with the candidate as it is now, including any edits you made to it in the Editor. It works only if the original's code hasn't changed since the candidate was made, or if it holds exactly the earlier candidate this one builds on (because you applied that one). Renaming the original or adding it to the Library doesn't count as a change. A strategy built without code can't be replaced: add the candidate to your Library instead. If the original is read-only, changing its code needs Trader or Quant.

The server can tell parameters, risk settings, timeframes and the rest of the code apart, but it can't tell entry and exit rules from filters inside the same function. When you allow one but not the other, the assistant follows your brief, but the server can't check it.

When generated code fails validation

Every generated strategy is checked before you can use it. If it still fails after the assistant's repair attempts, the reply shows Failed validation with the problems found. You can't save or run it; select Fix it to have the assistant repair it. A reply that writes no working strategy, or the same code again, is labelled too. None of these use AI credits (up to 10 a day). If the checker itself was unavailable, the code is labelled Couldn't be checked and you can still use it.

Bring your own code

AlphaProve Python: paste a strategy written for AlphaProve (it has HISTORY = … and def evaluate(ctx)), or select Open file to load a .py file. It's checked and opens as a new draft, without using AI credits. Importing your own Python needs Trader or Quant. To ask about code instead ("Why does this never trade?"), put the question above or below it: the assistant answers and nothing is imported.
TradingView Pine Script: see Pine Script.
Other platforms: MetaTrader (MQL4/MQL5), cTrader, NinjaTrader, backtrader, freqtrade, zipline, QuantConnect and EasyLanguage code can't be imported yet. Describe the strategy's rules instead and the assistant builds it.

Import a strategy from a video

Paste the link to a public trading video and the assistant reads it, lists the trading rules the speaker states, and can turn them into a strategy you can backtest. It works with:

YouTube: videos and Shorts (live streams and premieres can't be imported);
Instagram: Reels and posts;
TikTok: video links, and the share links the app gives you.

A chat that starts from a video link is titled Video import (YouTube), (TikTok) or (Instagram): the title names only the platform, never the link.

Paste the link on its own, or with a short note (up to about 200 characters). A message with more than one link, or a lot of other text, is treated as a normal chat message. Only public videos work: nothing signs in on your behalf, so a private or removed video can't be opened. While video imports are switched off, a pasted link is treated as an ordinary message.

Accept the import

The assistant answers straight away with an Import this video card. It names the platform and the link, and says what the import costs:

2 AI credits and 1 of this month's video imports. The import counts toward the month whether or not it works out.
Your plan's length limit.
When the credits come back (see what's refunded).

Nothing happens until you select Accept. Sending another message withdraws the card. If you are over the monthly limit, out of credits, or the platform is switched off for now, you get a plain reply instead of a card.

Plan	Video imports a month	Longest video
Free	3	10 minutes
Trader	30	30 minutes
Quant	100	60 minutes

The length limit depends only on how long the video is, not on the platform. You can run one import at a time; a second waits until the first finishes.

The speech is transcribed in English, Spanish, Portuguese, French, German, Italian or Romanian. A video in another language ends with a message saying so, and the credits are refunded.

What happens next

Accepting starts a task card that shows the video's title and length once they're known, and where the import is:

Stage	What's happening
Queued	Waiting for its turn. The card says Queued behind N with a rough wait, or Next in the queue. Shorts, Reels and TikToks go ahead of longer YouTube videos.
Downloading the video	Fetching the video, logged out.
Transcribing the audio	Turning the speech into text. A long video can take several minutes.
Reading the video	For clips up to 3 minutes, a few frames from the moments the speaker points at the screen are read as well, so rules shown only as on-screen text are caught. Longer videos use the speech only.
Finding the trading rules	Extracting the rules and checking them against what was said.

You can keep chatting or close the tab. Cancel import stops it.

What is refunded

The 2 credits come back if the import fails, if the video is longer than your plan allows, or if no trading rules are found. Refunds share the daily limit that failed generations use (10 a day). Past it, the card says the credits were charged.
Cancelling gives the credits back only while the import is still queued. Once it has started, they are charged.
The monthly import is used either way.
Build strategy (below) costs no credits. Generate code after it uses AI credits like any other generation.

If an import stops short, the card says why in plain words: the video is private or removed, needs a signed-in account, is a live stream, has no speech, is in an unsupported language, is too long or too large, or the platform is limiting automated access right now.

The report card

When the import finishes, the card shows what the video says, rule by rule:

The rule, the quote it comes from and a time link such as [4:12] that opens the video there (on YouTube; on TikTok and Instagram the link opens the video).
How we test it: Mapped (we can code it as stated), Approximated (we code our own reading, and the card shows the definition we used) or Dropped (we can't code it). Each rule also has a confidence of low, medium or high.
Assumed chips for anything the video doesn't say or says only loosely, for example Stop loss (not stated) or Timeframe: 15m (inferred). A number seen only on screen shows as on screen only.
Settings: market, timeframe, direction and sizing, each marked when it isn't from the video. With no market stated, the card uses BTCUSDT and says (default). If it names only an asset class ("crypto", "forex majors", "stocks"), the card uses that class's default market: BTCUSDT for crypto, EUR/USD for forex, SPY for stocks when we list it, and otherwise asks you to pick a market. If the video trades a market we have no data for, such as NQ, ES or US30 futures, the closest fund we do have (QQQ, SPY or DIA) is used and labelled Different market. A market the model leaves as "unspecified" (or similar) counts as no market stated. Any other market we have no data for isn't swapped for something else: the card is headed We don't have data for this video's market, names the market when it is a ticker (The video trades SPY, which we don't have data for yet), and shows a Market to build on picker. Build strategy and Fill the gaps with defaults stay disabled until you pick one, then build on it; the settings row reads AAPL (you picked it). Options and prediction markets can't be tested, and get no picker.
The video says: any results it claims, such as a win rate or a return, with the quote. The wording is ours, built from the claim's type and number; a number the video's quote doesn't back is left out.

If the video names several strategies, the card shows the first testable one and says how many there were.

Build strategy starts a new draft from the report, with no credits used, and the assistant continues as it does for any strategy: Generate code, check, add to your Library, backtest. The draft is written from the report's rules and settings, not from the video's own words.

A strategy built from a video is a normal strategy. Save it to your Library and backtest it on any market, timeframe or window you like, like any other strategy. Those backtests are ordinary backtests and get no claim check; the claim check below is a separate button.

When the video has no testable rules

Many videos talk up results or explain ideas without fixed rules. Then the card reads This video may not contain testable rules, gives the reason (hype, a signal-group ad, general education, a discretionary approach, lifestyle content or not about trading) and still shows anything it found. A video with no rules at all is refunded, and so is a video whose rules all use indicators we can't test yet (the card names them: Its rules use indicators we can't test yet: Ichimoku). That includes rules that need a line an indicator doesn't give us, such as the signal line of a linear regression slope. If some rules were found, the import is charged. You have three ways forward:

Fill the gaps with defaults builds a draft anyway, with the missing pieces filled in and marked as assumed (a 1.5 × ATR stop and a 2R target, for example). It appears when defaults are enough to make a draft. If the draft needed a default for something the strategy can't run without, such as a stop or an exit, the claim check can't judge the video's claims (see below). A swing take profit aims at the latest swing high (long) or swing low (short); a swing stop sits beyond it.
Describe the rules yourself puts a starter in the message box.
Paste the transcript puts a starter in the message box for you to paste the video's text after.

The claim check

Many videos claim a result: "80% win rate", "turned $1,000 into $10,000". The claim check tests those claims. It never starts by itself: you choose Check the video's claims on the video's report card. The button appears once the strategy you built from the video is saved to your Library (choose Build strategy, Generate code, then Add to Library). Until then the assistant tells you to save it first.

The button brings up a Backtest plan card, and nothing runs until you accept it. The plan runs the strategy as saved, on the video's own market and timeframe from the report card, with the settings the report showed:

Since the video was posted, up to today. This is a stretch the creator couldn't have picked after the fact.
The video's own period, as a second run, when the video's words give one (for example "2020 to 2023" or "the last 6 months") and we have data for all of it.

Each window is named on the card, with how it was cut. Free accounts get the last 89 days (last 89 days on Free), and a market's data may start after the video was posted.

Warm-up. Most strategies need history before their first signal (a 200-bar average, for example). The check loads that history from before the posting date, but nothing is traded or judged before it: no position, fee or equity change exists before the posting date. When the history you need doesn't fit in the data the plan can reach (your plan's lookback, or where the market's data starts), the judged start moves later and the card says why, for example last 89 days on Free; 11 days of warm-up first. The plan card and the receipt also say Indicators warm up on 1d data from 21 Mar 2025 (this strategy needs 250 bars before it trades); trades are judged from 2 Apr 2026. If no judged day is left, no plan card is made, and the reply says what the strategy needs and how much data is in reach. The video's own period is never moved: without room for its warm-up, it isn't run.

If you edit the strategy's code or ask for another market, timeframe or window, that is an ordinary backtest with no claim check.

The result carries a claim check under it:

The headline reads like Posted 2 Mar 2025. Since then: -6.1% vs +14% buy-and-hold over 212 trades.
A verdict for each claim, with the measured number beside it:
Holds up: the measured result matches or beats the claim (for a drawdown, stays within it). A win rate holds when the stated rate sits inside the 95% range of what we measured.
Did not hold up.
Too few trades: under 30 trades, which is too few to judge. The number is still shown.
Can't test: guarantees, "works on every market" and similar claims; a number the video didn't actually say; a return that the video gave for a different period than the one tested; a figure the video gives for another market (The video's figure is for TGT, not for the market we tested); or a result our data can't support.
What we assumed. The card lists what we filled in, such as Stop loss: 1.5 x ATR(14) (assumed). If Fill the gaps with defaults supplied something the strategy needs, a stop or an exit for example, every claim reads Can't test and says so: We had to assume the stop loss, so this tests our version, not the video's. The same goes for a market you picked (We tested on BTCUSDT, a market you picked: the video trades SPY.): every claim reads Can't test, except a claim about yet another market, which says so instead (The video's figure is for TGT, not for the market we tested).
Checked against the trades. Before a claim check is shown, the trades must match the strategy as written: the sides, stop distances and hold times. A run that fails these checks, or makes no trades, gets no claim check. The card also lists rules that were approximated or didn't make it into the code.
Forex. Our forex history has known bad prints, so a forex result never counts as Holds up; it reads Can't test with a note.
The cost line names the fees and slippage the backtest used, and the card is marked Hypothetical: a backtest isn't a promise of future results.

Under the result, buttons reuse what the assistant already does. Each sends a message that brings up its usual plan card, so nothing runs until you accept: Walk it forward, Would it pass a prop firm?, Test on other markets (for example BTC, EUR/USD and AAPL) for "works on every market" claims, Improve it, and Paper-trade it (on supported crypto perpetuals). Free accounts get Improve it only; the rest need Trader or Quant.

You can share a backtest that has a claim check, like any other backtest, from the chat or from the result's Share button. The Share dialog says when the link also shows the video's claims. The public page shows each claim, the verdict, the measured number (not for Can't test claims), what we assumed, the window, the cost line and Hypothetical. The creator's name is never shown.

The link to the video is off by default. On the result's Share dialog, tick Include the link to the video; in chat, ask for it to be included and the Share card says which you get. A share never includes a TikTok link, because it names the creator, and the checkbox doesn't appear for one; YouTube and Instagram links can be included.

What leaves AlphaProve

The video is downloaded and transcribed on our own servers. The transcript, the title and description, and the few frames read for a clip of 3 minutes or less are sent to our AI provider, DeepSeek, to extract the rules. Nothing else leaves. The video file is deleted when the import ends, and transcripts aren't written to logs. The public video's transcript and details are cached for up to 7 days, with no account information, so importing the same video again is quick.

Backtest and forward-walk plans

Ask the assistant to backtest one of your saved strategies, for example "Backtest my RSI strategy on BTCUSDT 1h for the last 6 months with 10,000 USDT at 1×." It answers with a Backtest plan card. Nothing runs until you accept it.

The card shows exactly what will run:

the strategy's name and a short revision label (the full revision is under Exact revision);
the market, timeframe and period, for example BTCUSDT · 1h · Mar 30, 2026 → Sep 29, 2026 · 10,000 USDT · 1×;
that leverage comes from the plan and replaces the strategy's own, while position sizing and other risk rules come from the strategy's RISK setting;
when the plan expires ("expires in 23 h"). Plans are valid for 24 hours.

Plans must fit your plan's limits (history length and maximum leverage) and the available market data; otherwise the assistant tells you what to change. A plan can't be edited. If you want different settings, ask for a new plan.

Accept and start starts the backtest at exactly those settings. It uses one backtest from your daily allowance, like any other run.
Cancel plan throws the plan away. It never stops a paper-trading run.

If the strategy's code or parameters change after the plan was made, or the plan expires, the card says so and Accept is disabled. Ask for a new plan. Renaming the strategy or adding it to your Library doesn't affect the plan.

If the strategy isn't in your Library, the card says it will be added; Accept adds it, so its runs show there.

A plan can also backtest a built-in strategy, and it can set the strategy's parameter values for that run. To test several strategies or symbols at once, ask for it in one message, for example "Backtest my three RSI strategies on BTC, ETH and SOL." One card lists every run (up to 5 saved strategies on up to 3 symbols) and the quota it uses.

Forward walks

Ask for a forward walk the same way, for example "Run a forward walk on my EMA strategy, BTCUSDT 1h, March to July, 2 months in-sample and 1 month out-of-sample." The assistant answers with a Forward-walk plan card. To test the strategy on the full window first, ask for a backtest and then a forward walk. You get one Backtest + forward-walk plan card with matching symbol, timeframe, window, capital and leverage, and one Accept and start runs both, in order. If the backtest fails, the forward walk doesn't start.

The strategy must declare parameters to sweep: a sweep list in PARAMS (see Parameters & forward-walk) or on a numeric no-code field. The walk uses that parameter grid, as on the Forward Walk page: up to 150 combinations on Trader and 200 on Quant.
Forward walks need Trader or Quant, and you can run one at a time. If another forward walk is queued or running, the step waits for a free slot (until the plan expires).
A walk can run for up to 1 hour on Trader and 3 hours on Quant. While it runs, the task card shows roughly how long is left; live batch progress is on the walk's own page.
Forward walks don't use your daily backtest allowance. In a combined plan, the backtest does.

Optimizer Lab campaigns

Ask for a Lab campaign by name, for example "Run a quick Optimizer Lab campaign on my EMA strategy, BTCUSDT 1h, from January 2025 to yesterday." Words like optimize or tune the parameters work too. Asking to improve a strategy makes a candidate instead (see above). The assistant answers with a Lab campaign plan card:

the parameters the Lab will tune and the ones it keeps fixed, and the estimated number of combinations;
the depth, its estimated duration, and its cost in AI credits: Quick 10, Standard 20, Deep 60;
how many Lab campaigns you have left this month;
a warning when earlier campaigns on this strategy already looked at the same final test period. A result that reuses it is not a test on unseen data.

Accept and start submits the campaign exactly as the Optimizer Lab page would, with the same plan limits, monthly quota and checks.

Lab campaigns need Trader (Quick and Standard, 8 a month) or Quant (all depths, 20 a month, up to 8 of them Deep).
The period must cover at least 600 days. The strategy needs 1 to 8 tunable parameters in PARAMS that give at least 50 combinations.
From chat, a no-code strategy stays no-code and the AI signal probe is off. Use the Optimizer Lab page for either.
You can run one Lab campaign at a time. If another one is running, the step waits for it, until the plan expires.
The credits are reserved when the campaign starts. If nothing clears validation, you get half back. A failed campaign is refunded in full. A cancelled one is charged only for the part that ran.

When the campaign ends, the task card shows the Lab's verdict and links the report. Ask AI to analyze explains it.

Compare two versions

To test a candidate against its original, or any two of your saved strategies, ask for a comparison and say what "better" means, for example "Compare the candidate with the original on BTCUSDT 1h for the last 6 months: higher Sharpe, and max drawdown must not get worse." The assistant answers with a Comparison plan card. It shows both strategies, the shared settings and the goal in words, such as Goal: higher Sharpe; max drawdown must not get worse.

Both strategies are backtested on identical market, timeframe, period, capital and leverage. The comparison uses two backtests from your daily allowance.
If you also ask for forward walks, both strategies get one. Their results are shown beside the verdict for information; they don't change it.
Position sizing still comes from each strategy's own RISK setting. If the two differ, the card and the result say so.
Accept adds both strategies to your Library, like any plan.

AlphaProve computes the verdict from the stored results. The AI doesn't choose it.

Verdict	When
Candidate better	The goal metric improved by more than the minimum difference, consistently across the period, and the metric you said must not get worse didn't.
Candidate worse	The goal metric got worse in the same way, or it improved but the metric you protected got worse ("improved Sharpe but worsened max drawdown").
Inconclusive	Fewer than 30 trades on either side, a difference below the minimum, or a difference that doesn't hold across the period.
Not comparable	A backtest didn't finish, or the two didn't run on the same settings, for example because a strategy was deleted or edited after you accepted.

The minimum differences are: - Sharpe 0.10; - return 1 percentage point; - max drawdown 1 percentage point; - profit factor 0.05.

For Sharpe, return and profit factor, the period is split into 12 equal slices and the improvement must hold across them. A max-drawdown goal uses the minimum difference only. Every result also says two things: - how many versions of this strategy you have tested: the more you try, the likelier one wins by luck; - that the period is not new data for a version designed after looking at it.

Changes you accept

On Trader and Quant, you can ask the assistant to change things in your account. It never makes the change itself: it shows a card that says, in plain words, exactly what will change. Nothing changes until you select Accept. Cancel plan throws the card away.

Ask for	What Accept does
Cancel a run	Stops a queued or running backtest, forward walk or Lab campaign at its next check. Results so far are kept.
Save the code	Saves the strategy code from this chat to your Library, with the name shown on the card.
Rename, describe, pin or unpin	Changes a saved strategy's name, description, or whether it's in your Library list. Unpinning keeps it saved.
Copy a strategy	Saves a copy, named on the card, to your Library.
Apply an improvement	Replaces the original strategy's code with an improved version's code. Its saved parameter values are cleared; its past runs are kept.
Delete a strategy	Deletes the strategy and its code from your account.
Delete a run	Deletes a backtest or forward walk and its results, with any share link to it. The strategy is kept.
Share a backtest	Makes a public link to the backtest's tearsheet, with the expiry and attribution shown on the card. The strategy's code isn't shared.
Turn off a share link	The public link stops working at once. The backtest is kept.
Change backtest defaults	Sets the symbol, timeframe, starting capital or leverage that new backtest forms start from.
Send feedback	Sends your bug report, idea or other feedback to the AlphaProve team, with the text shown on the card.

Cards that delete or publish something carry a warning and a red Accept:

Delete cards say This can't be undone.
Apply cards say This can't be undone. Check it before you accept.
Share cards say Anyone with the link can see what it shares.

If what the card names changed after it was made (the run finished, the strategy was edited, you sent a newer message), or the card expired, Accept is refused. Ask again for a new card. An accepted change shows as a task card with Open to the item it changed (none after a delete).

Task cards

Once accepted, the plan becomes a task card in the conversation. The task runs on our servers: you can keep chatting, close the tab, or come back later on another device and the card shows where it got to.

Status	Meaning
Starting	The task is about to start a run (after Accept or Retry).
Running	A run is being started (or its start is being retried after an interruption).
Waiting for results	A run is in progress. Results appear on the card when it finishes.
Waiting for capacity	All your run slots are busy; it starts when one frees up (until the plan expires).
Needs your input	Something changed since you accepted, such as the strategy's code or parameters were edited, your plan's limits changed, or the plan expired. Ask for a new plan or cancel.
Completed	The headline results are on the card (for a backtest: return, Sharpe, max drawdown).
Failed	The card says why in plain words.
Cancelled	You cancelled the task.

A forward-walk step shows Waiting for a free forward-walk slot while another walk is queued or running, then Running with about how many minutes are left. For a finished walk, the card shows its out-of-sample return, walk-forward efficiency, max drawdown and number of batches.

A Lab step shows Waiting for your other Lab campaign to finish while another campaign runs, then Running with a link to the live campaign. When the campaign ends, the card shows: - the verdict; - the number of trials and the checks passed; - AI credits used against the credits you accepted.

View report opens the Lab report. Cancel task cancels the campaign; its report so far is kept.

When a comparison finishes, its card shows both results side by side: each metric, the differences, the verdict and its reason, and links to both backtests.

View result opens the full result: a backtest's tearsheet in Results, a forward walk on the Forward Walk page. In a backtest + forward-walk task, each step has its own Result link.
Ask AI to analyze asks the assistant what the result means.
Cancel task stops further runs and asks its running backtest to stop. A backtest that is still loading market data can take a few minutes to stop. A running forward walk stops at its next batch, which can also take a few minutes. Results so far are kept, and paper-trading runs are never affected. If you cancel before the backtest starts running, it's given back to your daily allowance: straight away with Cancel task, or within a few minutes if you cancel it from Results. A backtest that already started, finished or failed still counts. Backtests you start outside a chat task (from Results, the Library or Run) keep their charge when cancelled while queued.
Retry appears only when a run was interrupted on our side (for example, a server restart). It starts one new run with the same settings and uses one more backtest from today's allowance. Nothing was charged twice for the interrupted run. For a forward walk, Retry starts one new walk with the same settings.
If our server stops in the middle of a forward walk, its step keeps showing Running until our clean-up marks it as interrupted. That can take up to about 3 hours 20 minutes after the walk started (the longest a walk may run, plus a margin). Then Retry appears.

Notifications

When a task finishes, fails or needs your input, a notification appears on whatever page you're on, naming the conversation and the result. On a computer it sits at the top right on chat and Terminal pages, and at the bottom right elsewhere. On a phone it appears at the top of the screen. Up to three show at once; the next one appears when you clear one.

Open chat takes you to the conversation. Dismiss clears it here, and on your other open tabs at their next check.
Notifications stay until you act on them. If the tab was in the background, they appear when you come back to it.
You don't get one for the conversation you're looking at, or for a task you cancelled yourself.
A task notifies once per outcome: if it fails again after Retry, there's no second notification.

Paper trading from chat

On Trader and Quant, you can ask the assistant to start, pause, resume or stop a paper run: a simulated run on live market data that never places real orders. See Paper Trading for how paper runs work. On Free, the assistant points you to the Paper Trading page instead.

To start one, say what you want, for example "Paper trade RSI Bounce on BTCUSDT 1h with $5,000". When your message says everything the run needs, it starts straight away. There is no plan to accept.

Strategy, market and starting capital: name them in your message. If you don't name the market, the assistant may reuse the market of your last backtest in this chat, otherwise it asks. Write the capital as money: $5,000, 5k USDT or 5000 USDT. A bare number isn't read as capital, except as your answer when the assistant asks how much capital to use. If you don't name a strategy, the assistant uses the one selected in the chat, or the one you backtested last in it.
Timeframe defaults to the timeframe of that strategy's last backtest in the conversation. If there isn't one, the assistant asks.
Leverage defaults to 1×. A leverage above your plan's maximum is refused, never lowered for you.
End: without an end date, the run continues until you stop it. An end date must be between 5 minutes and 90 days away.

If something is missing, the assistant asks, and your answer completes the request. If you give the timeframe, capital, leverage or end date twice, the later message wins. A bare ticker in a later answer doesn't replace a full pair like BTCUSDT you wrote first.

To pause, resume or stop a run, just say the action, for example "Stop my paper run". The assistant acts straight away only when exactly one of your runs can take that action. Otherwise it shows a run picker of up to 10 runs: the ones you mentioned first, then the ones started in this chat, then the newest. One click does it. Naming a market or strategy also brings up the picker, so you confirm which run you mean. You can pause only a running run and resume only a paused one. From chat, a run that's Catching up can be stopped but not paused or resumed. The card's Pause works like the Paper Trading page, so the card may offer it for such a run.

The paper run card shows the strategy, market, capital, leverage and end, with (default) next to the timeframe, leverage or end date when the assistant filled it in. Its status goes from Starting… to Running.

Open paper run opens the run on the Paper Trading page.
Pause and Stop ask the same confirmation as the Paper Trading page: open positions are closed at the last simulated price, and a stopped run can't be resumed.
If the run can't start, the card says why, for example when all your paper run slots are in use. A run that can't start within 10 minutes of your message never starts later: the card says That took too long to start — ask again.
Every card says Simulated — no real orders.

The assistant starts or changes a paper run only when your own message asks for it. A question, such as "Should I paper trade it?", gets an answer, not a run. Research, improvement and backtest requests, good results and text from web pages never start one. Each start counts toward your plan's paper limits, as on the Paper Trading page. Cancel task, Cancel plan and Stop response never pause or stop a paper run.

Stop and retry a reply

While the assistant is writing, Stop response ends that reply only. It doesn't cancel a task card. A stopped reply shows Response stopped. If it stopped before writing anything, Retry writes it again in place. For a partly written reply, use Regenerate under the reply.

History

Chat History lists every conversation in one place, including conversations from the former AI Trader. Old AI Trader links open the same conversation in AI Chat.

Filter by what happened in a thread: Research & analysis, Gather, Generate, Backtest review, Pine import, Pine export or Idle. Filtering searches older conversations too.
Pin, rename or delete a conversation. On a phone, these are under More actions.
Export as Markdown downloads a readable copy of the conversation: the messages, a short list of the tools used, the strategy and backtest analysis, and the sources with their dates. Images and links written by the assistant are exported as plain text.
Export with tool data downloads the same copy plus the raw tool inputs, outputs and structured results as JSON, in a separate file ending in "tool data".

What stays out of chat

Some things are only done on their own pages, never from chat. The assistant can open the page for you, but it can't change anything there:

billing and your plan;
deleting your account or exporting your account data;
your password and signing out of all devices;
agent connections (MCP): approving or removing an agent's access.

Limits

Answers use AI credits from your plan's monthly quota (your remaining credits are shown under Settings). Backtests, including accepted plans, use your daily backtest allowance. Forward walks don't, but you can run only one at a time.
A Lab campaign from chat uses AI credits like one started on the Optimizer Lab page, and counts toward the same monthly campaigns.
A video import uses 2 AI credits and one of your plan's monthly video imports; see what's refunded.
When the assistant only asks you to say what you mean, that reply is free (up to 20 a day). Generations that end with failed or no usable code are free too (up to 10 a day).
The assistant starts, pauses, resumes or stops a paper run only when your own message asks for it, and never places real orders. Backtest, forward-walk, Lab and comparison plans never touch a paper run.
AI analysis can be incomplete or wrong. Not financial advice.