Converta imagem para texto com IA em pt-BR. Use OCR online para transformar JPG, PNG e screenshots em texto puro ou formatado e baixar em TXT, Markdown ou HTML.

Scanlume is an OCR tool built to make image-to-text work more practical for Brazilian Portuguese users.

I started building it because a lot of OCR products feel either too generic or too heavy for everyday use. Some tools can extract text, but the result is often messy, overformatted, or difficult to reuse. In many real workflows, people are not trying to reconstruct a full document layout. They just want to upload an image, get readable text, and continue working without friction. That is the problem Scanlume is trying to solve.

The current version focuses on a very simple product idea: turn images into usable text in a cleaner and more flexible way. The tool works directly in the browser and is currently optimized for pt-BR-first usage. It is especially useful for screenshots, landing pages, banners, interface captures, lightweight documents, JPG files, PNG files, and image-based content that needs to be copied, reviewed, repurposed, or archived.

Scanlume currently offers two main modes.

The first is Simple OCR. This mode is designed for speed and low friction. It is the right choice when the user only needs plain text quickly. There is no extra formatting logic and no unnecessary complexity. It is useful for screenshots, posters, chat captures, notes, and any case where the main goal is just to extract the visible text fast.

The second is Formatted Text. This mode is meant for cases where preserving the main reading structure is more useful than returning a raw text block. Instead of trying to rebuild the original page pixel by pixel, it keeps the main hierarchy of headings and paragraphs so the result is easier to read and reuse. This is especially helpful when the extracted content is going into a document, a draft, internal notes, a CMS, or a content workflow. Users can export the result as TXT, Markdown, or HTML depending on what they want to do next.

One important part of the product is that the workflow stays lightweight. Users can upload multiple images, preview them first, remove individual files before running OCR, and then choose when to start processing. The interface is built to reduce unnecessary steps. There is also a progress view, a clearer mobile layout, and a simpler support flow inside the product.

The product is currently in a free trial stage. Anonymous trial usage is available, and users can also log in with Google to unlock a proper account experience with higher daily limits. An account page is already in place so the product can evolve from a simple trial tool into a more complete user system over time. There is also a waitlist for the future paid version, since the next phase of the product will include larger plans and billing support.

Under the hood, recognition is powered by a large multimodal model, but the product itself is intentionally positioned around usability, not around technical hype. The goal is not to overwhelm the user with AI language. The goal is to provide a clean OCR workflow that feels straightforward and useful in real work.

Scanlume is being developed incrementally, with a lot of attention on real usage instead of just feature lists. Current work includes better mobile conversion flow, cleaner account handling, support tools, waitlist management, and preparation for future paid plans. The paid version is not fully open yet, but the product is already structured in a way that can grow into subscriptions, larger quotas, and more repeat usage scenarios.

At this stage, Scanlume is best described as a practical OCR product for Brazilian Portuguese workflows: simple when plain text is enough, structured when readability matters, and lightweight enough to use without getting in the way.