Rolling your own serverless OCR in 40 lines of code
christopherkrapu.comNot sure what “your own” in the title is supposed to mean if you are running a model that you didn’t train using a framework that you didn’t write on a server that you don’t own.
I think in this case "your own" means under your control, rather than a service or license you pay for. "your own" as in ownership of artefacts, not as in being the creator.
Consider the source of the idiom: rolling your own cigarettes.
Which involves taking some rolling papers, a pouch of loose tobacco (or whatever), and perhaps a little device if you're rich. As opposed to manufactured cigarettes, you're just doing some manual assembly for the end-product.
You don't need to cultivate the plants or pulp any trees to roll your own.
I originally tried to do this on my own server but my GPU is too old :(
Slammed an A380 in my old server that doesn't even have a GPU power connector & it works pretty well for stuff that will fit on it. They're only like, $150 brand new nowadays; could be a decent option.
It means "one that is yours" in the same way "running your own plex server" does not imply starting with building a silicon fab.
<insert a random "i made this" meme>
And then call it serverless
Not sure what "baking your own bread" means if you are using wheat grown by someone else in an oven that you didn't build that is run with electricity you didn't created from your muscles' force. You haven't even contributed to the nuclear fusion which created the oxygen for the water molecules you've been using! How dare you, standing of the shoulders of giants!
Is it "building your own oven" if you go to Lowe's, buy an oven, and installed it yourself? You've done some work, but your integrating a pre-built appliance into your kitchen, not built your own oven
The title is "rolling your own", not "building your own"
This is more like buying a loaf of bread from the store and saying you baked it yourself. They did nothing even close to making an OCR engine.
why are chefs baking bread? there's buildings to construct
Wouldn't "Serverless OCR" mean something like running tesseract locally on your computer, rather than creating an AI framework and running it on a server?
Serverless means spinning compute resources up on demand in the cloud vs. running a server permanently.
~99.995% of the computing resources used on this are from somebody else's servers, running the LLM model.
> Serverless means spinning compute resources up on demand in the cloud vs. running a server permanently.
Not quite. Serverless means you can run a server permanently, but you need pay someone else to manage the infrastructure for you.
You might be conflating "cloud" with serverless. Serverless is where developers can focus on code, with little care of the infrastructure it runs on, and is pay-as-you-go.
> You might be conflating "cloud" with serverless. Serverless is where developers can focus on code, with little care of the infrastructure it runs on, and is pay-as-you-go.
That's not what serverless means at all. Most function-as-a-service offerings require developers to bother about infrastructure aspects, such as runtimes and even underlying OS.
They just don't bother about managing it. They deploy their code on their choice of infrastructure, and go on with their lives.
A runtime is notably NOT infrastructure, had you said instruction set you might have landed closer to making a compelling argument, but the whole point is that AWS (and other providers) abstract away the underlying infrastructure and allow the developers to as I said, have "little care of the infrastructure it runs on". There is often advanced networking that CAN be configured, as well as other infrastructure components developers can choose to configure.
Close. It means there's no persistent infra charges and you're charged on use. You dont run anything permanently.
It still doesn't capture the concept because, say, both AWS Lambda and EC2 can be run just for 5 minutes and only one of them is called serverless.
Unless the engineer takes steps to spin down EC2 infrastructure after execution, it is absolutely persistent compute that you're billed for whether you are doing actual processing or not. Whereas lambda and other services are billed only for execution time.
Depends if you mean "server" as in piece of metal (or vm), or as in "a daemon"
Thanks for noting this - for a moment I was excited.
You can still be excited! Recently, GLM-OCR was released, which is a relatively small OCR model (2.5 GB unquantized) that can run on CPU with good quality. I've been using it to digitize various hand-written notes and all my shopping receipts this week.
https://github.com/zai-org/GLM-OCR
(Shameless plug: I also maintain a simplified version of GLM-OCR without dependency on the transformers library, which makes it much easier to install: https://github.com/99991/Simple-GLM-OCR/)
When people mentions the number of lines of code, I've started to become suspicious. More often than not it's X number of lines, calling a massive library loading a large model, either locally or remote. We're just waiting for spinning up your entire company infrastructure in two lines of code, and then just being presented a Terraform shell script wrapper.
I do agree with the use of serverless though. I feel like we agree long ago that serverless just means that you're not spinning up a physical or virtual server, but simply ask some cloud infrastructure to run your code, without having to care about how it's run.
> When people mentions the number of lines of code, I've started to become suspicious.
Low LoC count is a telltale sign that the project adds little to no value. It's a claim that the project integrates third party services and/or modules, and does a little plumbing to tie things together.
>implement RSA with this one simple line of python!
No, that would be "Running OCR locally..."
'Serverless' has become a term of art: https://en.wikipedia.org/wiki/Serverless_computing
It's good they note explicitly:
> Serverless is a misnomer
Running it locally would typically be called “client(-)side”.
But this caught me for a bit as well. :-)
That's the beauty of such stupid terms.
I use carless transportation (taxis).
taxis are cars, aren't they?
Precisely. And serverless uses servers.
Yep. That fraudulent term finally got me this time. Totally serverless except for that remote 3rd party server. Sigh.
Deepseek OCR is no longer state of the art. There are much better open source OCR models available now.
ocrarena.ai maintains a leaderboard, and a number of other open source options like dots [1] or olmOCR [2] rank higher.
I wasn't aware of dots when I wrote the blog post. This is really good to know!! I would like to try again with some newer models.
you are comparing to DeepSeek's old OCR, there's DeepSeek-OCR2 which btw is amazing from my experimentations. https://huggingface.co/deepseek-ai/DeepSeek-OCR-2
The article mentions choosing the model for its ability to parse math well.
A bit surprised to learn that Rednote maintains one of the leading open-source OCR models on the market, nice.
hi. i run "ocr" with dmenu on linux, that triggers maim where i make a visual selection. a push notification shows the body (nice indicator of a whiff), but also it's on my clipboard
#!/usr/bin/env bash
# requires: tesseract-ocr imagemagick maim xsel
IMG=$(mktemp)
trap "rm $IMG*" EXIT
# --nodrag means click 2x
maim -s --nodrag --quality=10 $IMG.png
# should increase detection rate
mogrify -modulate 100,0 -resize 400% $IMG.png
tesseract $IMG.png $IMG &>/dev/null
cat $IMG.txt | xsel -bi
notify-send "Text copied" "$(cat $IMG.txt)"
exitI am working on a client project, originally built using Google Vision APIs, and then I realized Tesseract is so good. Like really good. Also, if PDF text is available, then pdftotext tools are awesome.
My client's usecase was specific to scanning medical reports but since there are thousands of labs in India which have slightly different formats, I built an LLM agent which works only after the pdf/image to text process - to double check the medical terminology. That too, only if our code cannot already process each text line through simple string/regex matches.
There are perhaps extremely efficient tools to do many of the work where we throw the problem at LLMs.
Slight tangent: i was wondering why DeepSeek would develop something like this. In the linked paper it says
> In production, DeepSeek-OCR can generate training data for LLMs/VLMs at a scale of 200k+ pages per day (a single A100-40G).
That... doesn't sound legal
HathiTrust (https://en.wikipedia.org/wiki/HathiTrust) has 6.7 millions of volumes in the public domain, in PDF from what I understand. That would be around a billion pages, if we consider a volume is ~200 pages. 5000 days to go through that with an A100-40G at 200k pages a day. That is one way to interpret what they say as being legal. I don't have any information on what happens at DeepSeek so I can't say if it's true or not.
Tried adding a receipt itemization feature into an app using OpenAI. It does 95% right but the remaining 5% are a mess. Mostly it mixes prices between items (Olive oil 0.99 while Banana 7.99). Is there some lightweight open source lib that can do this better?
So I'm trying to OCR 1000s of pages of old french dictionaries from the 1700s, has anything popped up that doesn't cost an arm and a leg, and works pretty decently?
Take a look at Mistral, https://mistral.ai/news/mistral-ocr-3
I use Gemini for that. Split the PDF into 50 page chunks, throw it into aistudio and ask it to convert it. A couple of 1000 pages can be done with the free tier.
Qwen3 VL.
Thanks! I'll have a look
How does this compare to Tesserect?
Question for the crowd -- with autoscaling, when a new pod is created it will still download the model right from huggingface?
I like to push everything into the image as much as I can. So in the image modal, I would run a command to trigger downloading the model. Then in the app just point to the locally downloaded model. So bigger image, but do not need to redownload on start up.
That book is freely available from its author in pdf format already… but I guess it’s about the journey?
If I had to guess, I would say that this method might be applicable to other books besides the one featured in the post.
I wanted to let an LLM be able to grep and read through it.
Always wondered how auth validation works on these. Could I use your serverless ocr?
Why "rolling"? Is this a reference to baking or what's the origin?
reference to cigarettes
The cold-boot time on this model can hardly be called “serverless”
The real question is why pick python 3.11
Uh... So I've been telling AI to write a single page html/js OCR app. And I'll include the pdf I want as an attachment.
I have 4 of these now, some are better than others. But all worked great.
tl'dr version:
step 1 draw a circle
step 2 import the rest of the owl