ShowHN Today 2024-09-08 en
1. Show HN: Velocity – Financial analysis at the speed of thought
URL: Visit Now
Author: blotterfyi
Text:
Points: 2 points
Comments: 1 comments
2. Show HN: How old do i look –– AI face age detector
URL: Visit Now
Author: letsgocn
Text: Upload your photo and use AI to estimate your age.
It’s a fun tool to see what your age might be in the eyes of AI.
Easy to use: No registration, no login, completely free.
Instant Results: Get your age estimation in seconds using AI.
User Privacy: Your photos are processed securely and never stored on our servers.
Points: 3 points
Comments: 8 comments
3. Show HN: Browserbased (Serverless Headless Browsers)
URL: Visit Now
Author: yevbar
Text:
Points: 2 points
Comments: 0 comments
4. Show HN: Ki Editor - the multicursor syntactical editor
URL: Visit Now
Author: hou32hou
Text: Hi everyone, I have been developing this editor, Ki, for over a year, and have employed it substantially in all kinds of development (including Ki itself) for at least 3 months.
I think it is mostly crystallized, thus I’m happy to share it with you today.
Its main strength is first-class multi-cursor and structural (syntax) editing, which is a rare combination in the realm of editors (TUI or GUI alike).
Hope you’ll enjoy it!
Points: 70 points
Comments: 20 comments
5. Show HN: Avoid Stripe Chargebacks with This
URL: Visit Now
Author: justdev
Text:
Points: 2 points
Comments: 0 comments
6. Show HN: Boiler Dev, is this a joke?
URL: Visit Now
Author: newstuffalways
Text:
Points: 3 points
Comments: 1 comments
7. Show HN: AI-Poems from New York Times Headlines
URL: Visit Now
Author: skaplich
Text:
Points: 1 points
Comments: 0 comments
8. Show HN: SFTP Bridge to S3
URL: Visit Now
Author: paulhenri_l
Text: Hey HN,
After seeing all the cool solopreneurs on X I decided to try and see for myself what this is all about. 9 months later here I am with my first project.
I decided to scratch my own itch, creating SFTP servers from a simple S3 bucket. I was tired of all my employer’s customers asking for SFTP access when all I wanted was to use S3. There I have my lifecycle rules, proper access control, lambda triggers. All the cool stuff. But they keep asking for SFTP and let’s be honest SFTP isn’t cool.
So I created this bridge, they get SFTP, I get modern tech. I hope this tool can help you feel something when using SFTP too. Would love your feedback.
Paul-Henri
Points: 7 points
Comments: 0 comments
9. Show HN: I made a webapp to create neat mandala pattern with interactive tools
URL: Visit Now
Author: Amithv
Text:
Points: 1 points
Comments: 1 comments
10. Show HN: I made a tool that helps devs easily deploy portfolio sites
URL: Visit Now
Author: esaminu
Text: I got tired of diving into old code and fixing responsiveness bugs just to keep my portfolio page up to date so I built LeetPost, a tool that lets you build and deploy stunning looking portfolios on unique personal domain names with just a few clicks.
Points: 6 points
Comments: 1 comments
11. Show HN: Reviewradar – an AI that helps you chat with software reviews
URL: Visit Now
Author: andreict
Text:
Points: 1 points
Comments: 0 comments
12. Show HN: Use Spectre.Console Terminal Widgets in PowerShell
URL: Visit Now
Author: BasiliusCarver
Text: I like spectre.console and wanted to be able to use it in powershell scripts so I built this module.
Points: 3 points
Comments: 2 comments
13. Show HN: Dump entire Git repos into a single file for LLM prompts
URL: Visit Now
Author: artkulak
Text: Hey! I wanted to share a tool I’ve been working on. It’s still very early and a work in progress, but I’ve found it incredibly helpful when working with Claude and OpenAI’s models.
What it does: I created a Python script that dumps your entire Git repository into a single file. This makes it much easier to use with Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) systems.
Key Features:
- Respects .gitignore patterns
- Generates a tree-like directory structure
- Includes file contents for all non-excluded files
- Customizable file type filtering
Why I find it useful for LLM/RAG:
- Full Context: It gives LLMs a complete picture of my project structure and implementation details.
- RAG-Ready: The dumped content serves as a great knowledge base for retrieval-augmented generation.
- Better Code Suggestions: LLMs seem to understand my project better and provide more accurate suggestions.
- Debugging Aid: When I ask for help with bugs, I can provide the full context easily.
How to use it: Example: python dump.py /path/to/your/repo output.txt .gitignore py js tsx
Again, it’s still a work in progress, but I’ve found it really helpful in my workflow with AI coding assistants (Claude/Openai). I’d love to hear your thoughts, suggestions, or if anyone else finds this useful!
https://github.com/artkulak/repo2file
P.S. If anyone wants to contribute or has ideas for improvement, I’m all ears!
Points: 17 points
Comments: 7 comments
14. Show HN: Traction dashboard for indiepreneurs to share in public
URL: Visit Now
Author: Bluemahogony
Text: Hey, I’m Denis, a Techstars alum and multi-founder.
After being inspired by indiepreneurs who share their progress publicly, I built a traction dashboard for startups to make ‘building in public’ easier by backing up their journey with transparent data.
Like many others, I’m also on my own journey of shipping more meaningful products, fostering transparency and community engagement.
It integrates seamlessly with Stripe and Google Analytics to automatically showcase your revenue and growth metrics in real-time.
You can set it up in under 5 minutes by cloning the GitHub repo. All credentials stay safely local, so you’re fully in control of your data.
I’d love any feedback, positive or negative — especially on what features would be most useful for indiepreneurs.
Points: 1 points
Comments: 0 comments
15. Show HN: I made an app for deciding with friends
URL: Visit Now
Author: algorr
Text: I made and publisb an mobile app that makes your decisions fun and social. Also it keeps your decisions in mind and you can check them. Also It in play store. App Store : https://apps.apple.com/tr/app/picked-make-decision/id6475396… Play Store : https://play.google.com/store/apps/details?id=com.appoint.co…
Points: 1 points
Comments: 0 comments
16. Show HN: An optimizing model router with easy BYOK and serverless inference
URL: Visit Now
Author: CuriouslyC
Text: Hey HN,
I’m building Key Trustee to simplify AI integration. Key Trustee is an optimizing model router (based on RouteLLM) that provides unified “bring your own key” functionality so that users only have to enter their keys once for all their enabled applications. Key Trustee is OAuth based, so you can re-use your application’s existing auth tokens, and it uses an OpenAI compatible API so it integrates with existing AI code easily.
Key Trustee also allows you to build secure “serverless” AI applications. As long as you have a working OAuth flow, your application can be built entirely in the client!
We’re planning to add a push-based payment system that integrates with Stripe, letting users approve or deny payments directly with micropayment support using API credits so you don’t need a separate in-app currency. It’s still in development, but we’ll keep you posted when it’s live.
We’re also planning to support pass through inference to local LLMs, so if you have an Ollama/Llama.cpp server running you’ll be able to use that with all your Key Trustee enabled applications.
My goal with Key Trustee is to make life easy for small and medium sized application developers. To that end, I’d love to hear your ideas, thoughts and feedback! My email is in my profile and I have a discord if you’d like to chat: https://discord.com/channels/1282344086430421060/12823440864…
Points: 1 points
Comments: 0 comments
17. Show HN: Dot IX: A diagram generator (2024)
URL: Visit Now
Author: azriel91
Text: Hey all, I made a diagram generator to solve common frustrations with diagramming tools.
You can think of this as GraphViz dot + Tailwind + easier markup.
The concerns I’ve catered for:
1. Free, no ads, no vendor lock in.
2. Offline (client-side) diagram generation.
3. Easily shareable diagrams via link, or SVG copied to clipboard.
4. No IP leakage by using the URL fragment.
5. Text input for easy updates of multiple nodes / edges.
6. YAML / structured input, so you don’t need a custom SDK / parser if you want to generate the data structure.
7. Interactivity is done using 100% CSS — no scripts, no foreign element, so they should render in places that accept SVGs.
And some links:
- Previous thread: https://news.ycombinator.com/item?id=36825069
- Repo: https://github.com/azriel91/dot_ix
- Presentation I did at a meetup: https://youtu.be/JpoG_O3hJss
- Slides: https://peace.mk/book/learning_material/dot_ix_diagram_gener…
Other things that some people may care about:
1. Embedded VS Code (monaco) + rendered SVG in your browser, so you don’t have to keep copy pasting between your text editor and your browser.
2. Interactively highlight relevant parts of the diagram, so you don’t have visual clutter. Think of those whiteboard diagramming sessions where you have to “go back to that part before”.
3. Styling is done via Tailwind classes, including animation.
4. It’s written in Rust, with the intention of having a programmatic API so diagrams can be generated real time. Though this is a longer term goal for a different project.
5. The input model is kinda stable, but if you use this for your diagrams, it’s best to save both the link (for future editing) + the output SVG (in case I change the input model). The compression is lz-string, though I half wonder if I should take compression away (easier to generate a link from other tools)
Points: 1 points
Comments: 0 comments
18. Show HN: Nexa SDK – Build powerful and efficient AI apps on edge devices
URL: Visit Now
Author: frednoodle
Text: Hey HN! Alex and Zack here from Nexa AI. We’re excited to share something we’ve been working on.
Our journey began with the Octopus series --- action models for mobile AI agents (https://huggingface.co/NexaAIDev/Octopus-v2). We focused on making sub-billion parameter models excel at function calling, making high accurate and fast function-calling possible on mobile and edge devices. But as we delved into developing full-fledged on-device applications, we hit a roadblock.
We realized that optimizing for function calling (tool-use) alone wasn’t enough. Building powerful on-device AI apps requires a diverse set of tools: language models with domain expertise, speech processing, image generation, embedding models and more. That’s when we decided to create Nexa SDK --- a comprehensive toolkit that brings together everything developers need to build powerful and efficient AI applications that run entirely on-device.
Here’s what Nexa SDK offers:
- Support for both ONNX and GGML models.
- An integrated conversion engine for making custom GGML Quantized Models for different device hardware requirements.
- An inference engine that supports language models, image generation models, TTS, audio generation models, and Vision-Language Models.
- An OpenAI-compatible API server with optimization in function calling.
- A Streamlit UI for rapid prototyping.
- An intuitive CLI for easy model management.
Backend optimizations for latency and power consumption on edge devices.
To showcase its capabilities, we’ve built several demo apps running entirely on your device (https://github.com/NexaAI/nexa-sdk/tree/main/examples):
- AI soulmate with uncensored model and audio-in/audio-out interaction.
A quick interface for uploading and chatting with PDFs like your personal finance documents.
A meeting transcription app supporting multiple languages and real-time translation.
But we’re just getting started! There are lots of exciting developments in our pipeline, and we can’t wait to share them with you soon!
Check it out: (https://github.com/NexaAI/nexa-sdk)
Docs: (https://docs.nexaai.com/)
If you’re excited about the future of on-device AI, we’d really appreciate your support. A star on our GitHub repo goes a long way in helping us reach more developers!
Cheers,
Alex & Zack
Points: 25 points
Comments: 4 comments
19. Show HN: Evolution of “who is hiring” vs. “who wants to be hired” over the years
URL: Visit Now
Author: Raed667
Text:
Points: 1 points
Comments: 0 comments
20. Show HN: Jsonb_apply: apply functions to Postgres JSON, like it’s Lisp
URL: Visit Now
Author: fforflo
Text: Hi HN,
Here’s a small Postgres extension I’ve been working on: jsonb_apply.
This extension allows you to dynamically apply functions to jsonb objects, both recursively and with optional arguments.
The general signature looks like this:
select jsonb_apply(doc jsonb, func text[, variadic “any” args1_n])
You don’t have to specify types for the arguments, but they’ll be used in lieu of parser hints, to find the appropriate function to call.
So you can do things like this.
select jsonb_apply(’{ “id”: 1, “name”: “John”, “messages”: [ “hello” ] }’, ‘replace’, ‘hello’, ‘bye’);
{“id”: 1, “name”: “John”, “messages”: [“bye”]}
Came in handy for the couple of cases I wanted it for.
It’s mostly text-oriented for now, but I’m putting it out there, just in case anyone else wants something more LISPy in Postgres.
Points: 3 points
Comments: 0 comments
21. Show HN: Terminal CLI AI Auto Complete Free
URL: Visit Now
Author: faangguyindia
Text: Many people pay for GitHub copilot. Here’s a way to get your own auto complete which is highly accurate and super fast, without any payment
Proof of concept.
Points: 2 points
Comments: 0 comments
22. Show HN: I built a Next 15 SaaS boilerplate that makes web app development easy
URL: Visit Now
Author: mjomaa
Text: Hi everyone,
how are you doing? Hope you also enjoy the summer weather :)
Most SaaS starter kits and boilerplates do not provide much in terms of web app functionalities.
The idea is to focus only on the web app part, providing a high quality, modern and flexible solution. Always go the extra mile, never take a shortcut.
Website: https://achromatic.dev
Demo: https://demo.achromatic.dev
The tech stack is
- Next.js 15, React 19
- Auth.js v5
- Prisma
- shadcn/ui
- nuqs, next-safe-action, nice-modal
- TypeScript
- Tailwind
Everything is implemented using RSC and server actions.
Points: 3 points
Comments: 2 comments
23. Show HN: Search San Francisco satellite imagery using natural language
URL: Visit Now
Author: furiousteabag
Text: Hey HN! We’re Alex and Szymon from Bluesight (https://bluesight.ai/), where we’re developing a foundation model for satellite data. We’ve created a demo to showcase the current capabilities of state-of-the-art models and identify areas for improvement.
Our demo allows you to search for objects in San Francisco using natural language. You can look for things like Tesla cars, dry patches, boats, and more.
Key features:
- Search using text or by selecting an object from the image as a source (“aim” icon)
- Toggle between object search (default) and tile search (“big” toggle, useful when contextual information matters, like tennis courts)
- Adjust results with downvotes (useful when results are water images)
- Click on tiles to locate them on a map
- Control the number of retrieved tiles with a slider
We use OpenAI’s CLIP model (https://openai.com/index/clip/) to put texts and images into the same embedding space. We do a similarity search within this space using text query or source image. We are using CLIP finetuned on pairs of satellite images and OpenStreetMap (https://www.openstreetmap.org/) tags (https://github.com/wangzhecheng/SkyScript) because vanilla clip performs poorly on satellite data. We pre-segment objects using Meta’s Segment Anything Model (https://segment-anything.com/) and pre-compute CLIP embeddings for each object.
We’d love to hear your thoughts! What worked well for you? Where did it fail? What features do you wish it had? Any real-world problems you think this could help with?
Points: 4 points
Comments: 0 comments
24. Show HN: Miles and Points on Google Flights
URL: Visit Now
Author: jackculpan
Text:
Points: 2 points
Comments: 0 comments
25. Show HN: ImmiGreat – Credit Karma for Canadian Immigration
URL: Visit Now
Author: bentice
Text: Hey HN,
ImmiGreat (immigreat.io), a mobile app designed to simplify the Canadian immigration process for DIYers who want more clarity and understanding of the process.
Immigration is a high-stakes and stressful. The government’s official UIs often make things worse, leaving people more unsure of themselves. Many resort to hiring expensive consultants or lawyers to navigate the system. ImmiGreat aims to solve this by giving users an intuitive, easy-to-use platform that helps them confidently manage the process on their own.
The ImmiGreat app and web tools do the following:
- Calculate your CRS score, which determines your eligibility for Express Entry, Canada’s express entry immigration program.
- Track Express Entry Draws to see the latest CRS cutoffs. We also provide notifications.
- Monitor and calculate scores for PNPs (Provincial Nominee Programs), which are mini provincial points systems that can help you get permanent residency faster.
ImmiGreat is currently bootstrapped and hope to generate revenue by partnering with banks, telcos, and other services that are relevant to newcomers in Canada. Newcomers are a huge growth market for many Canadian companies. The plan is to be like Credit Karma or ClearScore for Immigration. ImmiGreat will provide tailored recommendations for products and services to newcomers.
I’d love for the HN community to try it out and share feedback! No logins required.
Check it out at immigreat.io Or search for “ImmiGreat” in the app store.
Points: 3 points
Comments: 0 comments
26. Show HN: I made FeatureFlow, Ship features users (really) want
URL: Visit Now
Author: moviesofmood
Text: Hey HN
I like Canny, but let’s be honest—it’s too expensive ($99/month) and comes with too many features for most users.
So, I built FeatureFlow:
FeatureFlow helps you collect feedback from your customers, prioritize features, and build products your users will actually love. Don’t waste time building features that only 1% of your users will use.
Plus, you can leverage the most upvoted feature requests to guide your product roadmap!
Upcoming features: Integration with GitHub issue requests and many more coming soon!
Enjoy!
Points: 3 points
Comments: 0 comments
27. Show HN: I build a product for converting HTML to PDF batteries included
URL: Visit Now
Author: thomas-cogez
Text: Printerz is a powerful API that allows you to convert HTML documents into dynamic PDFs, featuring a templating system that easily adapts to any web technology. Whether you’re a developer or a business, our solution helps automate the generation of PDFs such as invoices, contracts, or reports with ease.
Why choose Printerz?
- Fast and simple API: Convert any HTML into PDF with great flexibility. Dynamic variables: Generate personalized documents based on your users’ data.
- HTML template hosting: Save time with our unique template storage and management system.
- Easy integration: Compatible with popular languages and frameworks, integrate in just minutes.
Ideal for:
Developers looking to automate PDF generation. Businesses aiming to unify and automate their document workflows. Marketplaces, billing services, or any platform requiring dynamic PDFs. Get started with Printerz today! Test our API for free today and revolutionize the way you manage your PDF documents.
Points: 3 points
Comments: 1 comments
28. Show HN: Graphtracks Beta
URL: Visit Now
Author: zft
Text: Graphtracks was born out of a personal need—while exploring decentralized social networks like Bluesky Social, I realized there were no analytics tools to track performance over time. As someone who loves diving into social media data, I knew there had to be a better way.
Graphtracks is here to solve that problem, offering powerful performance tracking for decentralized social media. With Graphtracks, you can monitor likes, follows, reposts, and replies, along with a simple metric we call “Total Engagement”—the sum of all four.
Reposts and replies are missing metric from bluesky itself, this is unique for graphtracks.
In the Beta version, our service is completely free and open to everyone!
Here’s what you can do with Graphtracks:
- Top 100 Bluesky Accounts: Check out the top accounts by engagement over the last hour, day, week, or month.
- Engagement Metrics: See the total number of likes, follows, reposts, and replies on a single page.
- Detailed Engagement Analytics: Dive into detailed engagement graphs for accounts in the top 100.
- Search Functionality: Look up any public Bluesky account and explore their engagement data.
Check us out and give graphtracks a try! If you’re active on Bluesky, see how your performance grows up over time and grow with us.
We’d love to hear your feedback—let us know what you think!
Points: 1 points
Comments: 1 comments
29. Show HN: Convert lecture notes to a personal YouTube video Course
URL: Visit Now
Author: MO-379
Text: Hello again,
I’m about to make a lot of changes to the main learning area and don’t want to make stupid decision, so thought to get ideas from here first.
Around 5 days ago I’ve made a post showing the first version of my app, the idea is to allowing the user to make their own outline and have my system create a youtube playlist.
Since them I made a lot of changes, now I would like to know how to improve the user learning experience for you, does the UI look inviting and feel comfortable to use ?, what should change ? what should stay ? what do you think ?.
Points: 5 points
Comments: 0 comments
30. Show HN: Vomitorium – all of your project in 1 text file
URL: Visit Now
Author: jwally
Text: Neat little nodejs/cli tool for putting all of your project’s files into 1. I built it for LLM assistance with small-ish projects.
Points: 6 points
Comments: 1 comments
31. Show HN: I Built an AI Recruiting Toolkit Without Subscription
URL: Visit Now
Author: kranzelbart
Text: Hey Hacker News community,
I’m Leon, and together with my co-founder Arun, we’ve been focused on making the recruitment process smarter and more efficient. Today, we’re launching the new Hiringtools AI, a one-time payment product designed to help you master the essentials of recruiting.
We’ve shifted from a credit-based SaaS model to offer something more practical: lifetime access to a comprehensive set of tools and resources that simplify recruiting. Here’s what you’ll get with Hiringtools AI:
- AI-Powered Recruitment Tools: Automate key tasks like creating job descriptions, filtering candidates, and improving job postings with AI.
- Prompt Library: Access a wide range of carefully crafted AI prompts tailored specifically for recruitment scenarios.
- Job Board Library: Explore a curated selection of job boards, helping you find the best platform for any role.
- Recruiting Services Library: Discover vetted recruiting tools and services to optimize your hiring workflow.
- Bonus: We’ve also included a Power-Prompt Ebook on employer branding, completely free, to help you craft stronger job listings and enhance your brand as an employer.
We’ve built Hiringtools AI for people who want to stay ahead in recruiting—whether you’re an HR professional, a startup founder, or just looking to optimize your hiring process. Our aim is to provide everything in one place, so you can focus on what really matters: finding the right talent.
Check it out: https://hiringtools.org
As always, we’re eager for feedback and suggestions on how we can make the product even better.
Points: 2 points
Comments: 4 comments
32. Show HN: Respeakable.com – A good AI Language Tutor
URL: Visit Now
Author: nicowaltz
Text: Hi HN,
We’ve just launched v1 of Respeakable, an AI-enhanced language tutor. We have two German lessons (Level A2, beginner to complete beginner) designed by a language-teacher friend of ours – feel free to try it out and give any feedback.
We (lschneider and I) have been experimenting with learning languages (specifically French) using different vocabulary apps and tools. The issue is that apps like Duolingo are after your attention (their motto is “test everything”) rather than teaching you a language and are therefore hyper gamified and too easy – all about hijacking your dopamine system. The only viable solution seems to be to attend real-life courses.
We always thought language learning is the perfect use-case for LLMs but we’ve been astounded at how bad all AI language tutors out there are, so we’ve decided to make a better one that we would actually use.
Respeakable uses structured lessons that are much like real-life lessons and only use LLMs for certain exercises, where it makes sense. We also use FSRS (free spaced repetition scheduler) and word frequency lists for practicing vocabulary on flashcards efficiently. FSRS has been shown to increase speed of memorization by ~30%. As a by-the-by: FSRS is also just being rolled out in Anki for those that know it.
We will be adding more lessons and features in the next few days so stay tuned.
Have a great day!
[1] FSRS: https://github.com/open-spaced-repetition/fsrs4anki/wiki/ABC…
Points: 4 points
Comments: 2 comments
33. Show HN: Subtle.sh – Zero-setup E2E encrypted messaging using OpenPGP and Soketi
URL: Visit Now
Author: ivstiv
Text: Hey HN,
I’ve been playing around with OpenPGP.js and Soketi lately and decided to build a simple E2E encrypted messaging app. It’s called subtle.sh. The idea is pretty straightforward: you need to send a secret to someone, but you don’t want to use Slack, email, or any other plain text medium. subtle.sh generates a fresh session with ephemeral keys, you share the link, and once the tab is closed, everything disappears.
Some key points:
- Zero setup required (no accounts, no installation)
- E2E encryption using OpenPGP.js
- Real-time messaging with Soketi (an open-source Pusher alternative)
- Built with Next.js, React, and Tailwind
- Self-hostable backend
You can take control of as much of the stack as you want:
1. Use the official deployment as-is for quick, ephemeral sharing
2. Self-host the entire stack for complete control
3. Host just the backend and connect it to the official frontend
It’s not meant to replace your password manager, but it could be handy when you need to quickly share credentials with someone you don’t have a shared vault with. The whole thing is open source, and I’ve included detailed docs on self-hosting, making it accessible even if you’re not a seasoned sysadmin.
I’d love to get your thoughts, feedback, or any security considerations I might have overlooked.
Website: https://subtle.sh
GitHub: https://github.com/ivstiv/subtle.sh
The reason I didn’t just use Pusher’s end to end capabilities was mostly because I wanted to Play with OpenPGP.js but also don’t plan on having Pusher as the sole communication mechanism in the future.
P.S. Yes, I know carrier pigeons are a viable alternative, but they’re not exactly zero-config and user-friendly.
Points: 4 points
Comments: 0 comments
34. Show HN: Node-scim / OpenAPI for SCIM v2
URL: Visit Now
Author: mnahkies
Text: I don’t normally share work at such a raw / early stage of development, but I thought I’d share this as:
- To my surprise I couldn’t find a decent openapi specification for SCIM v2
- I also couldn’t find any information about people using SCIM in conjunction with Firebase Auth / GCP identity platform
As I said this is incredibly raw / early - doesn’t support much, and the openapi specification is still extremely inaccurate, but in case anyone else is looking for a SCIM client example thought I’d share.
It also makes use of one of my other projects https://openapi-code-generator.nahkies.co.nz to generate most of the boilerplate from the openapi specification, and will hopefully turn into a good example of how to effectively leverage that.
(also if anyone sees this and can direct me to a accurate/complete openapi specification for SCIM v2 that would be incredibly helpful)
Points: 2 points
Comments: 0 comments
35. Show HN: Automate API Testing with Record and Replay
URL: Visit Now
Author: shailendraht
Text: This is shailendra here. Founder at HyperTest - hypertest.co
We are trying to make integration testing easy for developers. A lot of other teams and tools have taken a stab at this problem and having seen them we believe we have improvised the approach to help developers achieve this with minimum effort and pain.
How it works: Developers set-up our SDK (2-lines) in the source code their (backend) services and configure it to record traffic from any environment. When HyperTest works in RECORD mode it collects end to end trace of every incoming request i.e. the request, response and outbound calls.
These requests (tests) can be replayed on a new build later (pre-push or at CI) to check for regressions in API responses and outbound calls. In the REPLAY mode HyperTest uses mocked responses of all dependent systems to keep tests non-flakey and results deterministic and consistent. 3-min demo - https://www.youtube.com/watch?v=x6hmDUNFGW4
What does it do: HyperTest SDK auto-instruments all key functions and methods across all libraries you use to make outbound calls. This helps HyperTest mock these calls in REPLAY without asking developers to make any change in their source code.
How is this better:
- Set up is just like how you will set up an APM, i.e., 5 mins adding 2-lines of the SDK.
2. Support all protocols like http, graphQL, gRPC, Kafka and AMQP to cater to more use cases. Adding more as we speak
3. Test can be generated from any environment can be run anywhere even locally.
4. Active de-duplication to reduce the number of requests run on REPLAY. Optimise for code coverage & filter requests that don’t cover additional lines of code
5. Distributed tracing to help developers debug root cause faster
6. Auto-updates mocks as dependencies change to keep test results trustworthy.
HyperTest is currently available only for node projects. We work the teams with 5 or more services at the moment and have 50+ teams using it actively.
If this seems valuable can set-up a quick intro and explain how to get started here -https://calendly.com/shailendra-hypertest/30min
Would love feedback!
Points: 5 points
Comments: 4 comments
36. Show HN: Goroutine Monitor Powered by eBPF
URL: Visit Now
Author: keisku
Text:
Points: 5 points
Comments: 0 comments
37. Show HN: Open-source streaming API from ingest to player
URL: Visit Now
Author: matvp
Text: Hi there HN!
I’m Matthias and I’ve been into video streaming for quite a while now. Over the years I’ve built plenty of small setups, often CLI tools to test and debug streaming video. About a month ago, during my annual holiday, I started working on a simple API on top of my CLI tools. It’s time consuming to craft and run multiple ffmpeg jobs, get GOP / keyframes right, merge playlists, package to HLS, … One thing led to another and I eventually started exploring HLS interstitials, VMAP / VAST and dynamic ad insertion, MSE / EME (HLS.js in general), React (perf) as a player UI (although I have some experience with that from the past).
Once transcode or package jobs are finished, they’ll sync back to S3 (or any S3 compliant storage service).
From the get go I built in public, including:
1. Scale transcode (ffmpeg) jobs over multiple machines, crafting multiple video renditions, audio and text tracks. The API endpoint to start this job is rather simple, you define your source files (what you have, eg; an mp4 as video; an mp4 as audio with language “eng”, …) and the output (streams) of what you want (eg; H264 video 1080p, 720p, aac audio with language “eng”, …).
2. Package all transcode streams into an HLS playlist, uses shaka-packager under the hood. Uploads every playlist and segment to S3 publicly, at this point your asset is playable.
3. An on-the-fly playlist manipulator where you can limit the max. quality, add HLS interstitials for ad or bumper insertion or use VMAP / VAST to insert ads in stream. When a VAST is not yet transcoded and packaged, stitcher will start that job. When a playlist is found for that particular ad, stitcher will add an interstitial to the playlist. An HLS Interstitials supported player will then be able to play the ad at the given time defined in the VMAP.
4. A facade on top of HLS.js that makes working with interstitials a bit easier. Know that interstitials support with HLS.js is ongoing work (https://github.com/video-dev/hls.js/issues/5730).
5. When I was toying around, I felt I missed the experience of interacting with a player (that’s the only part in video streaming your end users interact with, and the default HTML5 player UI doesn’t fit more advanced use cases such as quality selection, display ad points on the seekbar, …) thus I built a small React component that functions as a player UI on top of the facade I built earlier.
I’m having a blast so far. Hope it provides value to some of you. Feedback is appreciated!
Points: 5 points
Comments: 0 comments
38. Show HN: I built my Website Design Agency
URL: Visit Now
Author: alexander2002
Text: Hi, Feel free to critique and give valuable advice.
Points: 4 points
Comments: 1 comments
39. Show HN: My Book DNA – Book recs based on real human connections (not AI)
URL: Visit Now
Author: bwb
Text: Hi all, creator here :)
I read a lot, and I want personalized book recommendations from other people based on my favorite books, authors, and genres. So, I created this tool using data we’ve been collecting for the last three years at Shepherd.com (which launched on HN three years ago).
The tool takes in 3 of your favorite books/authors and 3 of your favorite genres/topics.
Then, it shows you nine book lists based on your favorites to see if one resonates. Every book list on the website is made by an author or expert, so they either have passion or expertise (fiction or nonfiction).
And, you can sign up for a weekly email with new book ideas based on this. The email system is 100% personalized, so every person gets a 100% customized email every week—there is never a duplicate list.
Here is an example using my favorites:
https://shepherd.com/my-book-dna?r=books%3A34472&r=books%3A7…
I still have more to do here:
- Improve genre/topic accuracy; I am working on that this winter.
- I am working to launch a Book DNA format to try to decode why you loved a book and better match you with books based on similar readers.
- And generally improving this email as I get feedback :)
Shepherd is bootstrapped. I have a newsletter about building it and early access to new features here: https://build.shepherd.com/
What do we use to build this? Python, Django, Heroku, Postgre, Cloudflare, Postmark for email, NLP/ML for Wikipedia topic IDs via Wikifier (https://wikifier.org), Nielsen’s book API database (publisher data + Library of Congress data + BISAC), and Cloudinary.
My email is [email protected] if you want to share ideas or suggestions :)
Thanks, Ben
Points: 3 points
Comments: 0 comments
40. Show HN: Just released a new landing page for a blockchain usage data aggregator
URL: Visit Now
Author: jakemanger
Text: Would love to know what you think of it and whether there is any kind of data you would like to see!
Points: 4 points
Comments: 0 comments
41. Show HN: NODEFit – Fit Time-Series with Neural Differential Equations
URL: Visit Now
Author: gpavanb
Text: Hello everyone,
Time-series data is mostly based on observations around us and many physical models are based on differential equations.
However, accounting for non-ideal behaviour is difficult and using approaches such as regression does not capture the underlying dynamics.
Neural Differential Equations are an upcoming class of tools that allow for enhanced predictions by capturing the time-series not as a whole but by “fitting” the underlying dynamics.
Also included is the Stochastic Differential Equation model as well to capture dynamics in noisy situations.
NODEFit.jl is the faster Julia-based cousin that can be utilized if you are willing to live with the compile times.
GPU support is part of the PyTorch routines itself.
Points: 1 points
Comments: 0 comments
