How to pull prediction-market prices into Google Sheets

The short way

Write one small Apps Script function that fetches the price and writes it into a cell, and run it on a time-driven trigger. Do not reach for a formula first: IMPORTDATA reads only CSV or TSV, according to Google's own reference, and both venues' price endpoints answer in JSON.

Kalshi's market reads are public — the OpenAPI description of GET /markets/{ticker} carries an empty security list — and Polymarket's order-book prices need no key either. In the sheet, open Extensions → Apps Script and paste something like this:

// Writes the latest prices into the "Prices" tab. Column A holds the identifier.
function refreshPrices() {
  var sheet = SpreadsheetApp.getActive().getSheetByName('Prices');
  var rows = sheet.getRange('A2:A').getValues().filter(function (r) { return r[0]; });
  var out = rows.map(function (r) {
    var id = String(r[0]);
    if (id.length > 40) {
      // A long numeric id is a Polymarket outcome token: the midpoint of its book.
      var mid = UrlFetchApp.fetch('https://clob.polymarket.com/midpoint?token_id=' + id);
      return [Number(JSON.parse(mid.getContentText()).mid), new Date()];
    }
    // Otherwise treat it as a Kalshi market ticker: the last traded YES price.
    var res = UrlFetchApp.fetch('https://external-api.kalshi.com/trade-api/v2/markets/' + encodeURIComponent(id));
    return [Number(JSON.parse(res.getContentText()).market.last_price_dollars), new Date()];
  });
  sheet.getRange(2, 2, out.length, 2).setValues(out);
}

Then add a time-driven trigger for refreshPrices under Triggers in the editor. Write the timestamp beside every price, as above. It is the only way anyone reading the sheet can tell a live number from one that stopped updating on Tuesday.

What the options are

The Kalshi API is the cleanest source for this. GET /markets/{ticker} returns one market object with yes_bid_dollars, yes_ask_dollars, no_bid_dollars, no_ask_dollars and last_price_dollars, each a fixed-point dollar string with up to four decimal places, plus quantities as _fp strings. GET /markets?series_ticker= gives a whole series in one call, which matters once the sheet has more than a handful of rows.

The Polymarket CLOB API prices outcome tokens, not markets. Each market carries its token ids as a JSON-encoded array in a field called clobTokenIds, which you read once from Gamma and keep in the sheet. After that, /midpoint returns {"mid": "0.085"}, /price?side=BUY the lowest ask, /last-trade-price the last fill, and batch forms of all three take several tokens in one POST.

Apify's prediction-market actors are the no-code route that ends in an actual formula. An actor run on Apify's scheduler writes Kalshi and Polymarket rows — question, YES and NO price, bid, ask, volume — into a dataset, and the dataset items endpoint takes format=csv, which IMPORTDATA can read. Each actor is a different author's code with its own field names and its own per-row price, so pin one and expect to fix the sheet when it changes.

Artemis is the one card here with a real Sheets add-in, and it answers a different question. What it serves is venue-level daily metrics — volume, open interest, fees, users — across thirteen venues, not the price of a market. The add-in comes with the 100 USD Investor tier at 50,000 sheet calls a month, or alone as the 500 USD Plugin tier.

Lychee is for when the sheet is an analysis, not a monitor. Pull Kalshi or Polymarket history in the browser, export CSV or XLSX, and open the file in Sheets. There is no refresh; the plans start at 19.99 USD a month with 12,500 rows per historical pull.

Where this breaks

A cell is not live, and Sheets decides when it updates. Google's help page states the intervals: IMPORTDATA, IMPORTHTML, IMPORTFEED and IMPORTXML recalculate every hour, IMPORTRANGE every 30 minutes. A custom function — =KALSHI("TICKER") written as a script — is worse: Google's guide says it does not recalculate until you edit it or a cell passed to it changes, and passing NOW() to force it is not allowed and leaves the cell on "Loading..." for good. A custom function must also return within 30 seconds or show #ERROR!. That is why the short way above is a triggered script that writes values, not a formula.

The quota is per Google account, not per sheet. A consumer account gets 20,000 UrlFetchApp calls a day and 90 minutes of trigger runtime, a Workspace account 100,000 calls and six hours. A script that fetches 20 markets one at a time every minute makes 28,800 calls a day and stops part-way through the day. Batch: Kalshi's list endpoint by series, Polymarket's POST forms for many tokens.

The venue's rate limit sees Google, not you. Requests leave from Google's servers. Polymarket enforces its limits per IP address through Cloudflare and delays requests over them rather than rejecting them; Kalshi answers an over-limit request with a 429 and no Retry-After header. A script that hammers either one gets slow or empty answers, and the error lands in a cell you may not be looking at. Check the trigger's execution log.

Cents, dollars and probabilities are three different numbers. Kalshi now returns prices as fixed-point dollar strings — "0.1200" — and its own documentation says integer-cent fields cannot represent sub-cent ticks, which some markets now use. A tutorial dividing yes_bid by 100 was written against a field this endpoint no longer returns. Polymarket's prices are decimal strings between 0 and 1. Both are prices of a contract paying one unit, not probabilities: a wide spread or a stale last trade makes the "probability" in your sheet a quote of one side of a thin book. Both arrive as strings, so convert with Number() in the script rather than writing the string: a sheet whose locale writes decimals with a comma can keep "0.42" as text, and a sum over text cells does not fail — it leaves them out.

Which price is in the cell is a choice. Bid, ask, midpoint and last trade differ, sometimes by a lot, and the sheet does not say which unless you label the column. On a market that has not traded today, last_price_dollars is an old number that looks exactly like a current one.

Your own account is a different problem. Everything above is public data. Reading your Kalshi positions or balance needs signed headers — RSA-PSS over SHA-256, or an Ed25519 key — and Apps Script's built-in Utilities service offers RSA-SHA256 and HMAC signing but no PSS or Ed25519. Pasting a private key into a script bound to a shared sheet also gives it to every editor of that sheet. Treat it as a separate project; where your key lives is the page for that decision. The same goes for an Apify URL carrying token=: anyone who can read the formula can read the token.

If you outgrow this

A sheet polling once a minute is a dashboard with an hour's worth of failure modes. If what you want is a number that moves the moment the book does, that is a stream, not a spreadsheet, and streaming an order book is where that starts. If what you want is history to chart, pull it once from the price-history endpoints or an archive instead of keeping a sheet open — getting resolved market history covers the sources. And if the sheet is becoming a dashboard other people read, building a dashboard on Polymarket data covers the tools built for that job.