Daniel, your work is changing the world. More power to you.

I setup a pipeline for inference with OCR, full text search, embedding and summarization of land records dating back 1800s. All powered by the GGUF's you generate and llama.cpp. People are so excited that they can now search the records in multiple languages that a 1 minute wait to process the document seems nothing. Thank you!

▲ danielhanchen 2 days ago | parent | next [-]

Oh appreciate it!

Oh nice! That sounds fantastic! I hope Gemma-4 will make it even better! The small ones 2B and 4B are shockingly good haha!

	▲	qingcharles 5 hours ago \| parent [-]
		Just switched from 3.1 Flash Lite to Gemma-4 31B on the AI Studio API since there is a generous 1500/day on non-billed projects. It's doing fantastic.

▲ polishdude20 2 days ago | parent | prev | next [-]

Hey in really interested in your pipeline techniques. I've got some pdfs I need to get processed but processing them in the cloud with big providers requires redaction.

Wondering if a local model or a self hosted one would work just as well.

▲ evilelectron 2 days ago | parent | next [-]

I run llama.cpp with Qwen3-VL-8B-Instruct-Q4_K_S.gguf with mmproj-F16.gguf for OCR and translation. I also run llama.cpp with Qwen3-Embedding-0.6B-GGUF for embeddings. Drupal 11 with ai_provider_ollama and custom provider ai_provider_llama (heavily derived from ai_provider_ollama) with PostreSQL and pgvector.

People on site scan the documents and upload them for archival. The directory monitor looks for new files in the archive directories and once a new file is available, it is uploaded to Drupal. Once a new content is created in Drupal, Drupal triggers the translation and embedding process through llama.cpp. Qwen3-VL-8B is also used for chat and RAG. Client is familiar with Drupal and CMS in general and wanted to stay in a similar environment. If you are starting new I would recommend looking at docling.

▲

lwhi a day ago | parent [-]

Are you linking any of the processes using the Drupal AI module suite?

	▲	evilelectron 19 hours ago \| parent [-]
		Yes, they are all linked using Drupal's AI modules. I have an OpenCV application that removes the old paper look, enhances the contrast and fixes the orientation of the images before they hit llama.cpp for OCR and translation.

▲ chrisweekly 2 days ago | parent | prev | next [-]

Disclaimer: I'm an AI novice relative to many here. FWIW last wknd I spent a couple hours setting up self-hosted n8n with ollama and gemma3:4b [EDIT: not Qwen-3.5], using PDF content extraction for my PoC. 100% local workflow, no runtime dependency on cloud providers. I doubt it'd scale very well (macbook air m4, measly 16GB RAM), but it works as intended.

▲ patrickk a day ago | parent | next [-]

For those who wish to do OCR on photos, like receipts, or PDFs or anything really, Paperless-NGX works amazingly well and runs on a potato.

▲ polishdude20 2 days ago | parent | prev [-]

How do you extract the content? OCR? Pdf to text then feed into qwen?

I tried something similar where I needed a bunch of tables extracted from the pdf over like 40 pages. It was crazy slow on my MacBook and innacurate

▲ philipkglass 2 days ago | parent | next [-]

If you have a basic ARM MacBook, GLM-OCR is the best single model I have found for OCR with good table extraction/formatting. It's a compact 0.9b parameter model, so it'll run on systems with only 8 GB of RAM.

https://github.com/zai-org/GLM-OCR

Use mlx-vlm for inference:

https://github.com/zai-org/GLM-OCR/blob/main/examples/mlx-de...

Then you can run a single command to process your PDF:

  glmocr parse example.pdf

  Loading images: example.pdf
  Found 1 file(s)
  Starting Pipeline...
  Pipeline started!
  GLM-OCR initialized in self-hosted mode
  Using Pipeline (enable_layout=true)...

  === Parsing: example.pdf (1/1) ===

My test document contains scanned pages from a law textbook. It's two columns of text with a lot of footnotes. It took 60 seconds to process 5 pages on a MBP with M4 Max chip.

After it's done, you'll have a directory output/example/ that contains .md and .json files. The .md file will contain a markdown rendition of the complete document. The .json file will contain individual labeled regions from the document along with their transcriptions. If you get all the JSON objects with

  "label": "table"

from the JSON file, you can get an HTML-formatted table from each "content" section of these objects.

It might still be inaccurate -- I don't know how challenging your original tables are -- but it shouldn't be terribly slow. The tables it produced for me were good.

I have also built more complex work flows that use a mixture of OCR-specialized models and general purpose VLM models like Qwen 3.5, along with software to coordinate and reconcile operations, but GLM-OCR by itself is the best first thing to try locally.

▲

davidbjaffe 19 hours ago | parent | next [-]

Cool! For GLM-OCR, do you use "Option 2: Self-host with vLLM / SGLang" and in that case, am I correct that there is no internet connection involved and hence connection timeouts would be avoided entirely?

	▲	philipkglass 19 hours ago \| parent [-]
		When you self-host, there's still a client/server relationship between your self-hosted inference server and the client that manages the processing of individual pages. You can get timeouts depending on the configured timeouts, the speed of your inference server, and the complexity of the pages you're processing. But you can let the client retry and/or raise the initial timeout limit if you keep running into timeouts. That said, this is already a small and fast model when hosted via MLX on macOS. If you run the inference server with a recent NVidia GPU and vLLM on Linux it should be significantly faster. The big advantage with vLLM for OCR models is its continuous batching capability. Using other OCR models that I couldn't self-host on macOS, like DeepSeek 2 OCR or Chandra 2, vLLM gave dramatic throughput improvements on big documents via continuous batching if I process 8-10 pages at a time. This is with a single 4090 GPU.

▲

polishdude20 a day ago | parent | prev [-]

Thanks! Just tried it on a 40 page pdf. Seems to work for single images but the large pdf gives me connection timeouts

▲

philipkglass a day ago | parent [-]

I also get connection timeouts on larger documents, but it automatically retries and completes. All the pages are processed when I'm done. However, I'm using the Python client SDK for larger documents rather than the basic glmocr command line tool. I'm not sure if that makes a difference.

	▲	polishdude20 a day ago \| parent [-]
		Yeah looks like the cli also retries as well. I was able to get it working using a higher timeout.

▲ chrisweekly 2 days ago | parent | prev [-]

1. Correction: I'd planned to use Qwen-3.5 but ended up using gemma3:4b.

2. The n8n workflow passes a given binary pdf to gemma, which (based on a detailed prompt) analyzes it and produces JSON output.

See https://github.com/LinkedInLearning/build-with-ai-running-lo... if you want more details. :)

▲ tehologist a day ago | parent | prev | next [-]

Python pdftools to convert to images and tesseract to ocr them to text files. Fast free and can run on CPU.

▲ jorl17 2 days ago | parent | prev [-]

Seconded, would also love to hear your story if you would be willing

▲ Breza 19 hours ago | parent | prev | next [-]

I'm very active in family history and this kind of project is massively helpful, thank you

▲ irishcoffee 19 hours ago | parent | prev [-]

> your work is changing the world

I realize this may have been hyperbole, but it sure isn't changing the world.