Is Chessvision.ai Chess Position Scanner safe?
Chessvision is low risk. When you scan a chess position, the extension captures that part of your screen and sends the image to app.chessvision.ai over HTTP. The address uses http, so the image travels unencrypted.
Who publishes itSOFTWARE PAWEŁ KACPRZAK - no other listings under this identity, 4 shared hostnames
SOFTWARE PAWEŁ KACPRZAK - no other listings under this identity, 4 shared hostnames
Shared hosts - 4 hostnames
Hostnames hardcoded in this extension that few other listings call. That can mean one operator behind both, and it can equally mean a small shared vendor, so it is context rather than a conclusion. Hosts that many listings call are left out: they are services, not connections.
Chessvision findings
Chessvision.ai sends your screenshot over unencrypted HTTP
When you scan a chess position, the extension captures that part of your screen and sends the image to app.chessvision.ai over HTTP.
The address uses http, so the image travels unencrypted.
What actually happens
You ask the extension to scan a chess position.
The scan can cover the whole visible tab or a region you selected.
The extension turns that screen area into image data and sends it to Chessvision.ai over HTTP.
The configured prediction endpoint is http://app.chessvision.ai/predict.
Fields prepared for the prediction request
- Screenshot imagedata:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD2wBDAAYEBQYFBAYGBQYHBwYIChAKCgkJChQODwwQFxQYGBcUFhYaHSUfGhsjHBYWICwgIyYnKSopGR8tMC0oMCUoKSj (illustrative data URL prefix)
The visible chess-board image you asked to analyze. Nearby page content inside the selected area can be included too.
- Crop flagtrue
This tells the server whether you selected a smaller region instead of scanning the whole visible tab.
- Current playerwhite
This adds chess-position context to the image so the server can interpret the board from your side.
- Board orientationpredict
This asks the server to infer how the board is oriented in the captured image.
- Turn predictiontrue
This asks the server to infer whose turn it is from the scanned position.
Captured request
- Content-Type
- application/json
The code that does this
The scan image is wrapped in JSON and posted through an HTTP base URL
HTTP API base URL
background.bundle.jsconst config = { apiURL: "http://app.chessvision.ai", videoApiURL: "https://chessvision-video-search.appspot.com", apiKey: "AIzaSyAzopUSXAsZI9QwT4dSvl8PdUlf31Vd2gI", authDomain: "chessvision-scanner.firebaseapp.com", projectId: "chessvision-scanner", storageBucket: "chessvision-scanner.appspot.com", messagingSenderId: "473735200811", appId: "1:473735200811:web:9812e7fd521355e44bea6e", measurementId: "G-K7KRJ3TM98"};const apiBaseUrl = config.apiURL;Fetch wrapper used by prediction calls
background.bundle.jsfunction requestJson(url, options, respond) { return fetch(url, options) .then(function(response) { if (response.ok) { return response.json(); } var error = new Error("HTTP error: " + response.status); error.status = response.status; error.statusText = response.statusText; error.url = response.url; throw error; }) .then(function(data) { respond({ success: true, data: data }); }) .catch(function(error) { respond({ success: false, error: JSON.stringify(error, Object.getOwnPropertyNames(error)) }); });}Message handler posts the scan payload to /predict
background.bundle.jsif (message.action === "getFen") { var tabUrl = sender && sender.tab && sender.tab.url || ""; isForbiddenPage(tabUrl).then(function(result) { var forbidden = result[0]; var reason = result[1]; if (forbidden) { respond({ success: true, data: { success: false, forbidden: { reason: reason || forbiddenMessage } } }); } else { requestJson(apiBaseUrl + "/predict", { method: "POST", headers: { "Content-Type": "application/json" }, body: JSON.stringify(message.payload) }, respond); } });}Content script creates the image payload
contentScript.bundle.jscomponent.getFen = function() { var scannedUrl = window.location.href; var selectedRegion = component.state.scanMode === "select" ? component.selectPointsToCrop(component.selectPoints) : null; var ratio = window.devicePixelRatio; captureAndResize(selectedRegion === null ? null : { x: selectedRegion.x * ratio, y: selectedRegion.y * ratio, width: selectedRegion.width * ratio, height: selectedRegion.height * ratio }, function(imageDataUrl) { if (imageDataUrl !== null) { var payload = { image: imageDataUrl, cropped: selectedRegion !== null, current_player: "white", board_orientation: "predict", predict_turn: true }; chrome.runtime.sendMessage({ action: "getFen", payload: payload }, handlePredictionResponse); } });};Screenshot and canvas conversion helpers
contentScript.bundle.jsfunction getVisibleTabScreenshot() { return new Promise(function(resolve) { chrome.runtime.sendMessage({ action: "screenshot" }, function(response) { resolve(response.screenshot); }); });}function cropImage(sourceDataUrl, x, y, width, height, done) { var image = new Image(); image.onload = function() { var canvas = document.createElement("canvas"); var context = canvas.getContext("2d"); canvas.width = width; canvas.height = height; context.drawImage(image, x, y, width, height, 0, 0, width, height); done(canvas.toDataURL()); }; image.src = sourceDataUrl;}function resizeImage(sourceDataUrl, maxWidth, maxHeight, grayscale, done) { var image = new Image(); image.onload = function() { var scale = image.width > image.height ? maxWidth / image.width : maxHeight / image.height; scale = Math.min(scale, 1); var width = scale * image.width; var height = scale * image.height; var canvas = document.createElement("canvas"); var context = canvas.getContext("2d"); canvas.width = width; canvas.height = height; context.drawImage(image, 0, 0, width, height); if (grayscale === true) { var imageData = context.getImageData(0, 0, width, height); for (var row = 0; row < height; row++) { for (var column = 0; column < width; column++) { var offset = 4 * row * width + 4 * column; var gray = (imageData.data[offset] + imageData.data[offset + 1] + imageData.data[offset + 2]) / 3; imageData.data[offset] = gray; imageData.data[offset + 1] = gray; imageData.data[offset + 2] = gray; } } context.putImageData(imageData, 0, 0, 0, 0, imageData.width, imageData.height); } done(canvas.toDataURL("image/jpeg", 0.5)); }; image.src = sourceDataUrl;}Destination receiving the scan request
- app.chessvision.ai
Receives the chess-position prediction request at /predict using the extension's configured HTTP API base URL.
What Chessvision can do
Permissions this extension asks for, as declared in version 3.8.3. Asking for a permission is not a finding on its own - it is what the extension can do if it chooses to.
Read and change your data on app.chessvision.ai
*://app.chessvision.ai/*
Act on the current tab, but only after you click the extension
activeTab
Store data in your browser
storage
Run its own code inside the pages you visit
scripting
Chessvision version history
3.8.1
Mar 4, 2026
3.7.0
Mar 4, 2026
3.8.2
Mar 7, 2026
3.8.3
Apr 11, 2026
Chrome Web Store listing
- Versions
- 3.8.3 latest of 4 seen since March 4, 2026
- Store rating
- 4.6 from 752 ratings
- Last updated
- March 28, 2026
- Size
- 413 KiB
- Privacy policy
- chessvision.ai
