Skip to main content
Chess OCR with Chessvision.ai

A picture of a board.
A position you can use.

Chessvision.ai recognizes the pieces in a chess diagram and opens the position on an interactive board. Take a photo, choose a saved image, or share a screenshot to get started.

For iPhone and Android

See how recognition works
Chessvision.ai on an iPhone scanning a chess diagram from a printed book
A printed diagram becomes a digital position.
The idea behind chess OCR

What is chess OCR?

OCR is short for optical character recognition. It usually means reading letters and numbers from an image and converting them into usable text.

With a chess diagram, the useful information is the position: which piece is on each square, its color, and which squares are empty. That is the recognition task Chessvision.ai is built for.

Text OCRPrinted words → editable text

Recognize characters in a document.

Chess diagram recognitionBoard image → digital position

Recognize the arrangement of chess pieces.

How we turn an image into a position

Find the board. Read the squares.

Board detection and piece recognition solve two different parts of the problem.

  1. Locate the chessboard

    Find the board inside the picture and identify its grid of 64 squares. The diagram may be surrounded by book text or other content.

  2. Identify each square

    Determine whether a square is empty or contains a piece. For a piece, identify its type and whether it is White or Black.

  3. Open the position

    Represent the recognized position digitally and load it on an interactive board. Now you can move the pieces, analyze, or copy its FEN.

Example chess position with a White Knight on f5, a Black Rook on g5, and an empty fourth rank
A position from our original recognition article.

What is on this square?

Choose a square to see how its contents are described.

Square
f5
Piece
Knight
Color
White

NAn uppercase N represents a White Knight in FEN.

Piece-placement part of FEN for this example6k1/p2p2p1/2b4p/2K1PNr1/8/8/P2R4/8

This describes the pieces, rank by rank. A complete FEN also includes information such as whose turn it is and castling rights.

How we built the original recognizer. Our founder’s 2019 article describes a custom board detector and a neural network trained to recognize pieces, using both generated images and real examples.

Read the original technical story

Try it with the image you have.

The recognition happens inside Chessvision.ai. Choose the way that fits your position.

Take a photo

Point the app’s camera at a diagram in your book and take a picture.

Open an image

Select a saved chess image from Photos or Gallery inside the app.

Share a screenshot

Take a screenshot, tap Share, and send it to Chessvision.ai.

Recognition gets you to the board.

Try a move or turn on the engine to explore a continuation. Save a useful position with the star button and find it in Library’s Inbox with the same account.

Open Library
From the app stores

How players use the Scanner

If I can’t figure out why my variation didn’t work, I take a picture and get the answer within seconds.

LucidchessApp Store Chessvision.ai Chess Scanner · Review excerpt ·

No other programs accuracy comes close to this one on importing images.

Fanie NelGoogle Play Chessvision.ai Chess Scanner · Review excerpt ·
Also on your computer

Scan a diagram on a webpage.

The free browser extension recognizes chess positions from websites and opens them for analysis.

Chrome Web StoreChessvision.ai Chess Position Scanner4.5/5776 ratings100,000 users

Questions about chess OCR

What does chess OCR mean?
OCR stands for optical character recognition: turning text in an image into text a computer can use. Chess OCR is often used to describe recognizing a chess diagram. Chessvision.ai uses computer vision to turn the arrangement of pieces in that image into a digital position.
How is recognizing a diagram different from reading chess notation?
A diagram shows where the pieces are at one moment. Written notation records moves, such as Nf3 or Qxd4. The scanning workflow on this page recognizes diagram positions. A single board image does not contain the complete move history of a game.
Does recognition also find the best move?
Recognition opens the position on the board. Engine analysis is the next step: turn the engine on to see suggested moves, or leave it off and try your own ideas.
Can I scan a screenshot or a printed chess diagram?
Yes. Take a picture of a printed diagram with the mobile app, open an existing image from Photos or Gallery, or share a screenshot directly to Chessvision.ai.
Can I copy the recognized position as FEN?
Yes. Tap Share → FEN Text on the analysis board to copy or share the position. See the Image to FEN guide.
Can I use this with a PDF chess book?
Yes. The eBook Reader finds diagrams in PDF, ePub, and DjVu books. Open a diagram on a processed page to explore it on the board. Explore the Chess PDF Reader.
Can I try the scanner for free?
Yes. Free access includes limited scans. Membership includes unlimited scans. See plans.

Turn your next diagram into a board.

Get Chessvision.ai and explore the position behind the picture.

For iPhone and Android