AI Resources
Chandra OCR 2
Chandra OCR 2 is a document-reading model from Datalab. It takes page images and can reconstruct details such as selected checkboxes, table cells and handwritten text.
The companion tool accepts PDFs and images. Datalab publishes the model weights and examples you can compare with the original pages.
What it is
Read a document page
The model reads a rendered page as an image. Its companion code can produce Markdown and HTML from that output.
Why it stands out
Keep the mark with its label
Datalab’s form examples show reconstructed checkboxes alongside their labels. In a scanned questionnaire, a list of labels only tells you what the options were. The marks tell you which options the person selected.
Availability
Downloadable weights and companion code
The weights are available on Hugging Face under Datalab’s modified OpenRAIL-M terms. The companion code is published separately under Apache 2.0. Datalab’s model card lists the current terms for the weights.
Why it matters
What makes it useful
If you are turning a folder of scanned questionnaires into searchable records, you need to keep each answer with the question it belongs to. Start with a page that has several options and compare the extracted marks and labels with the scan.
What to know
Where it fits
Use the companion tool between your source files and the place you want to read the result. It accepts a PDF or image and can produce Markdown or HTML you can inspect before loading it into a document search tool.
Notable points
What stands out
A converted Markdown file can still contain HTML tables. The companion code keeps their HTML markup. Check that the reader or importer you plan to use displays those tables before sending a batch through it.
Before using
What to review
The repository documents a Transformers setup and a vLLM server setup for the downloadable model. Start with a non-sensitive sample page.
Compare checked and unchecked boxes, table headings and small print with the original. A clean-looking export can still contain a wrong answer.
Datalab describes its hosted service separately from the downloadable checkpoint. Evaluate the route you intend to use.
Check the original listing's documentation for current requirements, limitations, licensing and any relevant privacy or security considerations.
Reader fit
Who may find it relevant
This fits teams building a searchable collection from paper records. If the PDFs already have selectable text and you only need to search the words, try extracting that text first.
Editorial note
Why LifeHubber lists it
LifeHubber lists Chandra OCR 2 because its published form examples give you a way to evaluate the output your workflow will actually use. Try a form with two adjacent choices and only one checked, then look at that same output in your importer. If the marks survive extraction but disappear on import, you have learned where to fix the workflow before converting the rest.
Source links
Source materials
Reader note
About our listings
LifeHubber’s listings help you explore what’s available in AI and are not endorsements. We don’t independently verify every detail or guarantee that each listing is accurate, safe, actively maintained, or right for your needs.
More in AI Models
Keep browsing this category
Explore more AI model resources.
Gemma 4
google/gemma-4
A Google DeepMind Gemma 4 model family collection with public checkpoints including Gemma 4 12B, a dense multimodal model Google describes around local agentic workflows, native audio input, and encoder-free vision/audio handling.
Inkling 2 Small
thinkingmachines/Inkling-2-Small
An open-weight text and image model from Thinking Machines Lab with hosted access and two downloadable formats for multi-GPU deployments.
LFM2.5-350M
LiquidAI/LFM2.5-350M
A compact 350M-parameter model for on-device deployment, with 32K context and export paths for several local runtimes; Liquid AI recommends it for extraction, structured output, and tool use rather than knowledge-heavy or programming tasks.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
MORE ON LIFEHUBBER
Use AI on your own terms.
Explore more AI projects, stories, guides, tools and community insights across LifeHubber.