# JigsawStack Updates! Hackathon Highlights & What We're Building Toward 🧩

Hey Devs,

Things are moving fast. We launched JigsawStack v3, hosted another Smol Models Hackathon, shipped major improvements, and made our debut on Product Hunt. Here’s a quick recap, plus everything new and what’s coming next.

## Hackathon Recap

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1746228634862/2d6e0a3b-bd95-46a2-940f-28870e2d97ba.jpeg align="center")

We wrapped our second **Smol Models Hackathon** in San Francisco and it delivered.

Devs spent the day experimenting, connecting, and shipping creative prototypes using JigsawStack. From first-time builders to experienced founders, the energy was strong all the way through demo time.

Thank you to all who came through- we’ll see you at the next one soon!

## What’s New?

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1746819215075/3e03b549-d09a-408f-9662-ea5ebcaad8a3.png align="center")

v3 brings faster APIs, unified endpoints, and major pricing improvements. Here we highlight a few of our favorite changes in a fun way:

* [**AI Web Scraper**](https://jigsawstack.com/ai-web-scraper) that handles dynamic, static, and infinite scroll pages
    
* [**AI Web Search**](https://jigsawstack.com/ai-search) that finds anything, anywhere
    
* Token-based billing — **up to 36x cheaper**
    

[See it in action →](https://www.youtube.com/watch?v=YOjc4UHnqh4)

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1746820513893/6e6c8493-c643-4520-b3cd-c6624be7e7c9.png align="center")

### Gemini vs JigsawStack vOCR

We benchmarked **Google’s Gemini OCR** against **JigsawStack vOCR** on real-world tasks:

* Multilingual signs
    
* Structured PDFs
    
* Receipts
    
* Handwritten text
    

JigsawStack vOCR delivered stronger accuracy, richer structure, and faster responses.  
Try it out in our interactive [**Google Colab Notebook**.](https://colab.research.google.com/drive/1ShY7YsU7LXK5FHzoFShMom0sUS9_-nVa?usp=sharing)

[Read the full comparison →](https://jigsawstack.com/blog/gemini-vs-jigsawstack-ocr)

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1746820423800/d8ee8cb7-da47-43e9-a24e-540d3599ceb7.png align="center")

### Translate Text in Any Image

Our new [**Image Translation API**](https://jigsawstack.com/translation) lets you extract, translate, and overlay foreign text right back into images. While also keeping the images original style. Great for:

* Travel and tourism apps
    
* Product localization
    
* Educational tools and research
    

[Explore how it works →](https://jigsawstack.com/blog/translate-any-image)

### Other Improvements Since v3 Launch

* [**Speech-to-Text**](https://jigsawstack.com/speech-to-text): Faster speaker detection and cleaner transcript splits
    
* [**vOCR**](https://jigsawstack.com/vocr): Improved accuracy on high-res and complex documents
    
* **Bug Fixes**: Performance boosts across OCR, Translate, and Embedding
    
* **SDK Updates**: NSFW file support and more on the way
    

## Big Thank you to the Community!

Our **Product Hunt** launch brought in great feedback and support from devs around the world.

If you missed it, you can still check it out here:  
[View the launch →](https://www.producthunt.com/posts/jigsawstack-v3)

## Want to Get Involved?

Here are a few ways you can join in and maybe get featured in our next update:

* **Submit your build** → Reply to this email or share it on X. We’d love to feature you!
    
* **Join our next hackathon** → Coming soon (Bay Area)
    
* **Contribute to docs or SDKs** → Found a bug or missing example? PRs welcome 🙌
    

Thanks for building with us. Whether you're experimenting on the weekends or shipping to production, we're here to make backend AI infrastructure simple, powerful, and accessible.

Cheers,  
**The JigsawStack Team**  
support@jigsawstack.com | [Join our Discord →](https://discord.com/invite/dj8fMBpnqd)