Guide

How AlgoThesis works

Type a market idea in plain English; get back a cited, two-sided basket. Here's every piece — what it is, how to use it, and a worked example. New here? Start with The Wire.

The Wire (your home feed)

Just the news — with one line on how to play it.

The Wire is the home page: a running feed of real market news. What makes it different from a regular news site is that every story comes with three extra things — a one-line "how to play it" angle, a side-by-side Bull case vs Bear case, and a little 3-month chart of the tickers in the story. So instead of just reading what happened, you can see both ways the story could cut and turn any headline into your own thesis in a couple of clicks. You don't need an account to read it; you only sign in when you want to save a story or save a thesis draft.

How to use it

  1. 1Read top to bottom. The biggest story sits at the top as the lead, then stories run newest-first below it. Each card shows the topic and source, ticker chips with each stock's move today (green up / red down), the headline (click it for the original article), a short summary, and a news photo. A red left edge and a pulsing "Breaking" tag mark breaking news. The timestamp ("4 min ago") links to that story's own shareable page.
  2. 2Use the Bull vs Bear case to decide what YOU think. Under each story are two tinted panels — a green Bull case and a red Bear case — drawn from the filing/consensus behind the story. A small "The Angle" bar shows which way the evidence leans (or "Genuinely two-sided"). Click a panel to mark which side you lean toward; it's just for you (nothing is counted or shared) and it pre-fills your draft.
  3. 3See the numbers. Click "Show chart" on a card to open the basket mini-chart — the stocks in that story over the last 3 months, rebased to start at the same point, with a bold average line. On real-price charts you can hover to scrub and read the return at any point. The right-rail "Charts" toggle turns these on/off for the whole feed at once.
  4. 4Filter to what matters. The chip row has three views: "Everything" (default), "Breaking", and "Affects your book" (only stories touching a ticker you've saved). The number on each chip tells you how many stories qualify. "Topics ▾" narrows to one area — Semis, Earnings, Macro, Energy, Crypto, FX, or Geopolitics. Filtering is instant — the page never reloads. "Load more stories" pulls in older stories.
  5. 5Keep an eye on the right rail: the market strip, the most-mentioned tickers right now (Trending), and "On deck" — upcoming earnings/events marked BMO (before open) or AMC (after close). Signed in with saved tickers, "On deck" personalizes to your book.
  6. 6Turn a headline into a thesis. Either click "Draft thesis →" on any story — it bundles the bull/bear case + tickers into a prompt and saves it to your Dashboard to finish later — or use the "What's your thesis?" box pinned at the top of the feed: type a stock, sector, or idea (or pick an "Ideas" starter) and hit enter to jump into the builder.
  7. 7Returning visitors get a heads-up: a banner says "N new stories since you were here," and if fresh stories land while you read, an "↑ N new stories" pill appears at the top. Power move: press / to jump to search, J / K to step through stories, Esc to clear.

Example

The lead story is "Nvidia guides Q4 above estimates on data-center demand." The chips show NVDA and AMD up. You weigh the two panels — Bull ("backlog widening, pricing power intact") vs Bear ("already priced in, China export risk") — and click "Show chart" to see NVDA, AMD and the basket average over 3 months. You lean bullish, so you click the green Bull panel; the button now reads "Draft bullish thesis →". You click it, it saves the bull/bear context to your Dashboard, you keep scrolling, and later you open the Dashboard and finish that draft into a full cited, cited, two-sided basket.

When to use it: Start here every session — it's the fastest way to catch up on the market and turn something you just read into a thesis without hunting for a ticker first.

Thesis Builder

Type a market hunch in plain English, get back a basket — tickers, weights, reasons, sources.

Thesis Builder turns a plain-English market idea into a concrete basket of stocks: each one with a long/short tag, a percentage weight, and a one-line reason it's in there. It's a chat, not a one-shot search box — once you get a basket back you can keep asking it to change things ("swap out the unprofitable names," "make it more defensive") and it rebuilds. It exists so you don't have to translate a hunch into a list of tickers yourself, or go hunting for which companies actually have exposure to a theme.

How to use it

  1. 1Open New Thesis. You'll see "What's your thesis?" and a box that reads "Ask about a stock, sector, or type a thesis...". Type your idea and hit the arrow (or Enter).
  2. 2No idea yet? Click a quick-pick chip (AI infrastructure, Nuclear energy, Defense plays, Top gainers today, Earnings this week, GLP-1 pharma, Cheap tech stocks…) or open "Browse idea library." Each fires a thesis instantly.
  3. 3Watch the basket stream in: a short read on your thesis up top, then an "Example basket" card listing each ticker with a long/short tag, a weight %, and a one-line reason. If it searched the web, citation links appear under the answer as "Sources."
  4. 4Read the basket: the ticker is clickable for company fundamentals, the weight % is its share of the basket, and below you may see a value chain, a "Contrarian View," and a "What would make this wrong" box. Click "Show chart" for a hypothetical backtest line vs the S&P 500.
  5. 5Refine it in the chat. Use the "Ask a follow-up…" bar or the "Go deeper" chips: tighten the list, add a hedge, focus on cheaper names. It edits the same basket instead of starting over. "+ another variant" gives a different angle on the same thesis.
  6. 6Keep what's useful: signed-in generations stay in History. Choose "Review & track," inspect the exact snapshot, then confirm it to the dashboard only when you want to freeze the basket's dated EOD price baseline and begin separate basket-performance and evidence-health records. "Code" exports a runnable Python backtest script (copy or download .py).

Example

You think AI data-center power demand will strain the grid. A vague "AI stocks" gets a generic list — so type the causal chain instead: "AI data centers are going to blow up electricity demand and the grid isn't ready — who benefits?" The agent reads it back ("the bottleneck isn't chips, it's power"), maps the value chain, and returns a basket like VST 25%, GEV 25%, ETN 20%, CEG 20%, VRT 10% — each with its own reason and a "what would make this wrong" note. Then you follow up: "drop anything already up more than 80% this year and add a utilities ETF as ballast." It rebuilds with your constraint. Happy with it? Open "Review & track" and confirm the snapshot to start a live record, or choose "Code" to download the backtest and run the numbers yourself.

When to use it: Whenever you have a market view but not a list of tickers — a theme, a macro call, a "who actually benefits from X" question — and want a weighted, reasoned basket you can interrogate and refine.

Screener

Sort the whole stock universe by the fundamentals that matter to you, then hand the winners straight to a thesis.

The Screener is a sortable table of companies ranked by their reported numbers: revenue, year-over-year growth, profit margins, return on equity (ROE), return on invested capital (ROIC), debt, and free-cash-flow margin. The figures come from companies' own SEC filings, so it covers fundamentals only — not live prices or chart signals. Use it to go from "I want fast-growing, high-margin companies" to an actual short list in a few clicks, then check the ones you like and build a thesis on them.

How to use it

  1. 1Open the Screener from the left nav. It loads as one big table; the count above it (e.g. "480 results") tells you how many companies you're looking at.
  2. 2Click a preset filter tab to narrow the list: All, High Growth (revenue +20% YoY), Compounders (3-yr growth >15%), High Margin (net margin >20%), High ROE (>20%), High ROIC (>15%), Cash Machines (FCF margin >15%), Low Debt (debt/equity <0.5), and Profitable (any positive net margin). One tab is active at a time; the count updates instantly.
  3. 3Use the search box to jump to a company by ticker (NVDA) or name. Search stacks on top of whichever tab is active.
  4. 4Read the table left to right: ticker, company, sector, then the numbers. Color is a quick read, not a verdict — green = strong, red = negative or high-debt, gray = middling or missing. A dash (—) means the company didn't report that figure.
  5. 5Sort by any number by clicking its column header (▼ highest-first, ▲ lowest-first; click again to flip). Sorting reorders the entire filtered list; 30 rows show per page (Prev / Next at the bottom). Click any ticker for its full stock page.
  6. 6To act on it, tick the checkbox on one or more rows — a "Build basket with N →" button appears top-right; click it to carry those tickers into a new thesis. To keep a filter combo, name it and save it as a screen (Free keeps up to 3; Pro is unlimited).

Example

You want fast-growing companies that actually make money. Open the Screener, click the High Growth tab (revenue +20% YoY), then click the ROE column header so ▼ shows highest-first. You scan down: a row with Rev Y/Y 34% (green), Net M 28% (green), ROE 41% (green) is exactly the profile you wanted. You tick its checkbox plus two similar rows; the "Build basket with 3 →" button appears and drops you into a new thesis pre-loaded with those tickers. Before leaving, you save the filter as "Profitable high-growth" so it's waiting next time.

When to use it: When you don't have a specific stock in mind yet and want to surface candidates by the numbers — fundamentals first, before researching any single name.

Dashboard & Track Record

Theses you choose to track, with price performance and evidence health kept separate.

The Dashboard is your research desk for generated baskets, Wire drafts, and the theses you explicitly track. Tracking freezes the basket, weights, the latest available licensed end-of-day closes, and the request time. From then on, AlgoThesis reports two different things: hypothetical basket performance from those dated EOD baselines, and thesis health based on later supporting, weakening, or invalidating evidence. The baseline is not an intraday fill. A price gain does not prove the reasoning, and contradictory evidence does not rewrite the return.

How to use it

  1. 1Open the Dashboard from the nav. Overview summarizes tracked-basket performance; History remains the place to reopen generated research that you have not chosen to track.
  2. 2Start deliberately: open a generated basket, choose "Review & track," and confirm the exact snapshot to the dashboard. Confirmation freezes its positions and the latest available licensed EOD closes, whose actual market dates are shown separately from the request time. Generating or reopening a thesis alone does not start tracking. Refining it does not start tracking either.
  3. 3Read basket performance as a market result, not a truth score. Positive Rate, Avg Return, and best return compare the frozen EOD baselines with the latest available EOD closes; days tracked uses the separate tracking-request time. They are hypothetical and do not represent an actual position or intraday fill. AlgoThesis does not infer a price path or benchmark comparison when it only has an endpoint return.
  4. 4Read Thesis Health separately. New stories are classified against the original reasoning as supporting, weakening, invalidating, or irrelevant; the health score summarizes that evidence. A dash means there is not enough classified evidence yet.
  5. 5Use the Saved tab to revisit saved research and active records or stop tracking one. Stopping does not turn the last return into a verdict on the thesis, and it does not erase the original starting snapshot or evidence history.

Example

Three weeks ago you reviewed and confirmed "AI chip demand will triple by 2027," with NVDA, AMD, and TSM. Today its basket performance is +8.4% from the dated EOD baseline shown on the record, while Thesis Health is 72/100 because two later stories supported the demand case and one weakened it. Those numbers answer different questions: where the basket moved, and whether new evidence agrees with the original reasoning. Neither says that you acted on it.

When to use it: When you want to compare what the market did after you started tracking with what later evidence says about the reasoning itself.

Watchlist & AI Analyst

Keep your tickers in the sidebar, and let the Analyst write a fresh thesis on the names you've published theses on.

The Watchlist is the list of tickers in your left sidebar — each row shows the last close and the day's change, and clicking a ticker opens its full stock page. The Watchlist AI Analyst is a separate Pro feature that writes a short thesis for each ticker in your published theses — a one-line headline, a direction read (long, short, or neutral), the reasoning, and the catalysts to watch. It runs automatically every Monday at 7am ET and emails you the report; on Pro you can also run it on demand any time.

How to use it

  1. 1Find your Watchlist in the left sidebar (if it's collapsed to a thin rail, click the arrow at the top to expand). New accounts start with AAPL, NVDA, GOOGL, and BTCUSD.
  2. 2Add a ticker: click the + next to "Watchlist," type a symbol (TSLA), press Enter. For crypto, type BTC, ETH, or SOL. Each row then shows last close + day change; click the symbol for its stock page.
  3. 3Remove a ticker: hover its row and click the small x next to the price. Your list is saved in this browser, so it stays put on this device.
  4. 4For the Analyst, first publish a thesis on the names you want covered — it reports on tickers from your published theses, not your whole sidebar. No published theses yet? The Analyst page shows a "Publish a thesis first" button.
  5. 5Open the Analyst: on Pro, an "Analyst" link appears in the sidebar nav. Each card shows the ticker, a LONG / SHORT / NEUTRAL read, a headline, the thesis, and any catalysts.
  6. 6Refresh on demand (Pro): click RUN NOW at the top-right to generate a fresh report immediately instead of waiting for the Monday 7am ET run. On Free, the Analyst page shows an upgrade screen.

Example

On Pro, you want the Analyst to cover Tesla. You expand the sidebar, click + next to "Watchlist," type TSLA, hit Enter — a TSLA row appears (last close 248.50, -1.20%). You click it, open its stock page, and publish a thesis on it. Then you open the Analyst link and click RUN NOW. A TSLA card appears: a green "▲ LONG" tag, headline "Delivery beat eases demand worries," a two-sentence thesis, and a CATALYSTS list ("Q3 delivery report," "Cybertruck ramp"). The same report lands in your inbox, and from then on refreshes every Monday at 7am ET.

When to use it: Use the Watchlist to keep an eye on a handful of tickers and their daily moves; use the Analyst when you want a written, catalyst-aware read on the names you've actually published theses on.

Free vs Pro: which plan you need

Start free, stay free as long as it fits, upgrade the day a limit gets in your way.

AlgoThesis has two plans: Free and Pro ($29/mo). They both do the same core thing — type a plain-English idea, get back a researched basket of tickers with the reasoning. They differ only in how much you can do per day. Free is built to be genuinely usable on its own (no credit card), so upgrading is about removing daily limits once you're using AlgoThesis a lot — not about unlocking the basics.

How to use it

  1. 1Start on Free. No card. You get 20 thesis generations a day, 30 AI chat messages a day, up to 3 saved screens, and up to 3 alerts. Every thesis returns up to 3 strategies. Enough to decide if it fits your workflow.
  2. 2Watch which limit you actually hit. Each cap is independent: generations and chat reset every 24 hours; saved screens and alerts are running totals (3 of each). Never bump into them? Stay on Free — it's not a trial and doesn't expire.
  3. 3If the daily caps are what's slowing you down, go Pro ($29/mo): unlimited thesis generations, unlimited AI chat and follow-ups, and unlimited saved screens. It also turns the Watchlist Analyst from a cached take into one that runs live on demand, and bumps alerts to 50.
  4. 4Upgrade from the Pricing page ("Start Pro") or the "Why Pro?" page. Plan changes take effect right away.
  5. 5Cancel anytime from account settings. You keep your paid plan until the end of the billing period, then drop back to Free — your saved work stays.

Example

On Free, you research solar: "US solar names that could benefit if rates fall" returns a weighted, reasoned basket — generation 1 of your 20 for the day. Three follow-up questions in chat use 3 of your 30 daily messages. You save the screen — 1 of 3 slots. Over the next hour you run 6 more theses and save 2 more screens: now 7 of 20 generations, all 3 saved-screen slots used. The next time you try to save a fourth screen, Free stops you — that's the moment to consider Pro, where saved screens are unlimited and the daily caps go away.

When to use it: When you're not sure whether Free is enough or what the $29 upgrade actually changes — the answer is decided by which daily limit you keep hitting.

Ready to try it?

Type your first thesis — it's free, no card.

Write a thesis →