OCR — Extraindo Texto de Imagens e Documentos Digitalizados | Extracting Text from Images and Scanned Documents
-
OCR
PT-BR | EN
O que e
OCR (Optical Character Recognition) e a tecnologia que extrai texto de imagens — como fotos, prints de tela, documentos digitalizados e PDFs escaneados.
Seu agente consegue ler qualquer texto visivel em imagens e converter para texto editavel.
OCR vs Vision
Recurso OCR Vision Foco Texto puro Contexto visual completo Extrai Caracteres, palavras, paragrafos Descricao de cenas, objetos, pessoas Ideal para Documentos, notas, placas, codigos Analise de imagens, perguntas sobre cenas Comando ocr_image vision_tool (action=analyze) Exemplos Praticos
Extrair Texto de um Print
- Extraia o texto deste screenshot
- Leia o que esta escrito neste print
Documentos Digitalizados
- Extraia o texto deste PDF escaneado
- Leia os dados desta nota fiscal
Placase e Sinalizacao
- Leia o texto desta placa
- Extraia as informacoes do cartaz
Formatos Suportados
Formato Extensao Observacao PNG .png Ideal para screenshots e graficos JPG .jpg, .jpeg Fotos e imagens comprimidas TIFF .tiff, .tif Documentos digitalizados de alta qualidade BMP .bmp Imagens bitmap sem compressao GIF .gif Imagens animadas (le o primeiro frame) PDF .pdf Documentos multipagina ou escaneados WEBP .webp Imagens web modernas Dicas para Melhores Resultados
- Use imagens nitidas com boa resolucao
- Garanta bom contraste entre texto e fundo
- Evite fotos tremidas ou com reflexo
- Para documentos escaneados, prefira 300 DPI ou mais
- Texto impresso tem melhor resultado que texto manuscrito
Dica: Combine OCR com Desktop Automation! Primeiro capture a tela (screenshot), depois extraia o texto (ocr_image).
What It Is
OCR (Optical Character Recognition) is the technology that extracts text from images — like photos, screenshots, scanned documents, and PDF files.
Your agent can read any visible text in images and convert it to editable text.
OCR vs Vision
Feature OCR Vision Focus Pure text Full visual context Extracts Characters, words, paragraphs Scene description, objects, people Best for Documents, notes, signs, codes Image analysis, questions about scenes Command ocr_image vision_tool (action=analyze) Practical Examples
Extract Text from a Screenshot
- Extract the text from this screenshot
- Read what is written in this print
Scanned Documents
- Extract text from this scanned PDF
- Read the data from this invoice
Signs
- Read the text on this sign
- Extract the information from the poster
Supported Formats
Format Extension Notes PNG .png Great for screenshots and graphics JPG .jpg, .jpeg Photos and compressed images TIFF .tiff, .tif High-quality scanned documents BMP .bmp Uncompressed bitmap images GIF .gif Animated images (reads first frame) PDF .pdf Multi-page or scanned documents WEBP .webp Modern web images Tips for Best Results
- Use sharp images with good resolution
- Ensure good contrast between text and background
- Avoid blurry photos or reflections
- For scanned documents, prefer 300 DPI or higher
- Printed text gives better results than handwriting
Tip: Combine OCR with Desktop Automation! First capture the screen (screenshot), then extract the text (ocr_image).
-
R Rodrigo Serpa moved this topic from Getting Started on
-
R Rodrigo Serpa moved this topic from Copy & Paste on
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login