xx123122
Show HN: OpenTalking – Privately deployable real-time AI avatar
Show HN: InkSight – An open-source, LLM-powered e-ink display for "slow info"
https://github.com/datascale-ai/inksight
Generate personalized content based on the current environment (weather, time, date, solar terms) through backend LLMs (DeepSeek / Tongyi Qianwen / Kimi) and display it on a 4.2-inch e-ink screen.
I’m sharing an open-source project out of our lab called InkSight (code-named inco). Like many of us, our team found ourselves constantly distracted by notifications, glowing monitors, and endless doomscrolling on our phones. We wanted a way to consume high-quality, low-frequency information (like a Stoic quote, a minimalist daily briefing, or a quick recipe) without the cognitive load of a traditional screen.
So, we built InkSight—an open-source "slow tech" desktop companion. It uses an ESP32-C3 and an e-ink display to fetch customized LLM-generated content.
The Tech Stack & Architecture:
Hardware: ESP32-C3 written in C/C++ (Arduino framework). It supports common 2.13" / 1.54" e-ink panels.
Backend: Python & FastAPI. It acts as the brain, parsing user-defined JSON prompt templates and calling any OpenAI-compatible LLM (OpenAI, DeepSeek, Kimi, etc.).
Web Dashboard: Pure HTML/JS/CSS for easy configuration without digging into the code.
What we think HN might find interesting (The Optimizations): Instead of the ESP32 waking up, making an API call to the LLM, and waiting 5-10 seconds for a response (which drains the LiPo battery significantly), we built a smart caching layer in the backend. The backend pre-generates and caches the content. When the ESP32 wakes up, it fetches the cached payload in sub-seconds, updates the e-ink screen, and immediately goes back to deep sleep. This allows it to run for 3 to 6 months on a single charge and stay resilient against temporary network drops.
Deployment: We wanted to make it accessible to everyone, so the backend can be 1-click deployed to Vercel (for free), and the hardware uses a Captive Portal for zero-code WiFi and API configuration. Of course, it’s completely self-hostable if you prefer to run it locally.
Repo: https://github.com/datascale-ai/inksight
I’d love to hear your thoughts on the "slow tech" philosophy, any feedback on the architecture, or ideas for new content templates! I'll be hanging around the thread to answer any questions.
Show HN: InkSight – An open-source, LLM-powered e-ink display for "slow info"
https://github.com/datascale-ai/inksight
Generate personalized content based on the current environment (weather, time, date, and solar terms) via backend LLMs (DeepSeek / Qwen / Kimi) for display on a 4.2-inch E-ink screen.
My English is not good, so I use GPT to help translate and polish my replies to be polite. Maybe it made them sound too robotic. I am reading every comment myself. Sorry for the wrong impression.
We are actually three first-year Master's students. This project is indeed a summary of our learning from this past semester, which we rushed to wrap up right before the Chinese New Year break.
When I mentioned 'Project Lead,' I was referring to a senior PhD candidate in our lab. He acts as a mentor to review our code and ensure quality control, but the learning and implementation are very much ours. And yes, to move fast and polish the English, we did utilize LLMs during the writing process.
Thanks for understanding, and Happy New Year!
We are pleasantly surprised by the warm reception. We know the project (and our English localization) is still a Work in Progress, but we are committed to improving it to meet the high standards of the HN community. We'll keep shipping updates!
I'm currently a Master's student at USTC (University of Science and Technology of China). I've been diving deep into Data Engineering, especially in the context of Large Language Models (LLMs).
The Problem: I found that learning resources for modern data engineering are often fragmented and scattered across hundreds of medium articles or disjointed tutorials. It's hard to piece everything together into a coherent system.
The Solution: I decided to open-source my learning notes and build them into a structured book. My goal is to help developers fast-track their learning curve.
Key Features:
LLM-Centric: Focuses on data pipelines specifically designed for LLM training and RAG systems.
Scenario-Based: Instead of just listing tools, I compare different methods/architectures based on specific business scenarios (e.g., "When to use Vector DB vs. Keyword Search").
Hands-on Projects: Includes full code for real-world implementations, not just "Hello World" examples.
This is a work in progress, and I'm treating it as "Book-as-Code". I would love to hear your feedback on the roadmap or any "anti-patterns" I might have included!
Check it out:
Online: https://datascale-ai.github.io/data_engineering_book/
GitHub: https://github.com/datascale-ai/data_engineering_book