Show HN Daily Selection (2025-06-27)
- Magnitude – Open-source AI browser automation framework
- I built an AI dataset generator
- What time is it in Corporate
- Built a Food Scanner for Longevity
- I built a JSON-RPC library for Zig with compile time reflection
- Chat with ChatGPT, Claude, Grok, Gemini and Llama on One UI
- Rust ↦ WASM, K-Means Color Quantization Crate for Image-to-Pixel-Art
- Inworld TTS – high-quality, affordable, and low-latency TTS
- Pocket2Linkding – Migrate from Mozilla Pocket to Linkding
- AI Phone Interviewer – get a call in 30 seconds
- Anytype – a local and collaborative database with API and MCP server
- An open-source app to query 10 AI models at once
- Piper-mode – Text-to-speech for Emacs using the Piper TTS engine
- Tinder-style trip planning – Trip planning for the brainrot era
- Use Apple Container with Gemini CLI
- Voice-Mode MCP – Conversational Coding for Claude Code, Gemini CLI
Magnitude – Open-source AI browser automation framework [Developer Tools]
37 points, 14 comments
Open-source AI browser automation framework using a vision-first approach for reliable web interactions.
- Vision-first approach avoids flaky DOM navigation, improving performance on complex interactions.
- Supports drag-and-drop, data visualizations, legacy apps, canvas/webGL-heavy sites, and remote desktops.
- Uses visually grounded models (Claude Sonnet 4 or Qwen-2.5-VL 72B) for precise pixel-based actions.
- Fine-grained control with
act()andextract()syntax, allowing custom code integration. - Enables high-level task automation, low-level actions, and intelligent data extraction with zod schemas.
- Easy setup via "npx create-magnitude-app" for quick experimentation.
Discussion Link, Original Link
I built an AI dataset generator [Developer Tools]
115 points, 22 comments
AI dataset generator tool.
- Generates datasets for AI training
- Simplifies data collection process
- Supports various data types and formats
Discussion Link, Original Link
What time is it in Corporate [Business]
5 points, 0 comments
A visualization tool for corporate time formats, including the National Retail Federation 4-5-4 calendar and other requested calendar types.
- Visualizes the National Retail Federation 4-5-4 calendar for clarity
- Supports multiple corporate calendar types
- Plans to integrate SEC dataset for company symbol search (not yet implemented)
Discussion Link, Original Link
Built a Food Scanner for Longevity [Health]
3 points, 1 comments
A food scanner app that rates products based on their impact on longevity factors like inflammation and toxins.
- Scans food products to assess longevity impact
- Evaluates factors like inflammation, omega ratio, and toxins
- Currently in private beta for feedback
- Focused on health span and deep nutrition
Discussion Link, Original Link
I built a JSON-RPC library for Zig with compile time reflection [Developer Tools]
6 points, 0 comments
A JSON-RPC library for Zig using compile-time reflection to enable dynamic function dispatching in a statically typed language.
- Uses Zig's comptime feature for compile-time reflection to determine function parameter and return types
- Packages functions into uniformly typed call objects for dynamic dispatch
- Solves the challenge of handling functions with varying parameter counts and types in a static type system
- Provides a clean interface pattern for managing function calls in a map
Discussion Link, Original Link
Chat with ChatGPT, Claude, Grok, Gemini and Llama on One UI [ChatBots]
4 points, 3 comments
A unified interface to chat with multiple AI models and compare their responses.
- Supports ChatGPT, Claude, Grok, Gemini, and Llama in one UI
- Enables side-by-side comparison of responses from different models
- Helps users identify the best AI model for specific queries or research
- Streamlines research by eliminating the need to switch between platforms
Discussion Link, Original Link
Rust ↦ WASM, K-Means Color Quantization Crate for Image-to-Pixel-Art [Image]
6 points, 0 comments
Rust to WASM crate for converting images to pixel art using K-means color quantization.
- Converts images to pixel art
- Uses K-means color quantization
- Written in Rust and compiled to WASM
Discussion Link, Original Link
Inworld TTS – high-quality, affordable, and low-latency TTS [Audio]
18 points, 11 comments
Inworld TTS offers high-quality, affordable, and low-latency text-to-speech with multilingual support and emotional markup.
- Two models: TTS-1 (small, SOTA quality) and TTS-1-Max (larger, 3.5% better WER)
- Supports 11 languages and emotional markup (e.g., "[happy]")
- Built on LLaMA 1B/8B backbones, trained with GRPO and Xcodec2 neural audio codec
- ~500ms p90 latency for streaming API (TTS-1)
- Simple pricing at $5/1M characters
- Training and benchmarking code to be open-sourced on GitHub
- Optimized serving with Modular and Mojo-written MAX server
Discussion Link, Original Link
Pocket2Linkding – Migrate from Mozilla Pocket to Linkding [Productivity]
3 points, 0 comments
A tool to migrate bookmarks from Mozilla Pocket to the self-hosted Linkding service.
- Facilitates transition due to Mozilla Pocket's upcoming shutdown
- Supports self-hosted Linkding as an alternative bookmarking solution
- Plans for future enhancements beyond Linkding's current features
- Export service remains available until October 8, 2025
Discussion Link, Original Link
AI Phone Interviewer – get a call in 30 seconds [Business]
14 points, 7 comments
AI Phone Interviewer provides instant AI-powered screening calls for hiring processes.
- Get called in 30 seconds for a 2–3 minute interview
- Handles general screening questions and generates reports
- Plans to add technical screening, ATS integrations, and multi-language support
- Targets recruiters and startup founders for feedback
- Voice quality designed to feel natural
- MVP stage, seeking validation for future development
Anytype – a local and collaborative database with API and MCP server [Productivity]
14 points, 0 comments
Anytype is a local-first, privacy-focused collaborative database and wiki tool with end-to-end encryption and peer-to-peer sync.
- Local API for workflow automation (desktop-only currently)
- MCP server for LLM integration
- Raycast extension example
- CRDT-based sync for collaboration
- User-owned keys for authentication
- Open-source core (part MIT licensed)
- Supports docs, notes, tasks, tables, and media
- Real-time collaboration across users and devices
- Web publishing from desktop
- Native mobile companion apps
Discussion Link, Original Link
An open-source app to query 10 AI models at once [Developer Tools]
4 points, 2 comments
An open-source app that queries 10+ AI models simultaneously and summarizes their responses using a combiner AI model.
- Supports parallel queries across multiple AI models (Gemini, Claude, etc.)
- Summarizes responses for easier comparison
- Open-source and available for local use
- Includes a demo GIF and detailed technical blog post
Discussion Link, Original Link
Piper-mode – Text-to-speech for Emacs using the Piper TTS engine [Audio]
3 points, 1 comments
Emacs text-to-speech mode using the Piper TTS engine.
- Integrates Piper TTS with Emacs
- Enables text-to-speech functionality within Emacs
- Lightweight and efficient implementation
Discussion Link, Original Link
Tinder-style trip planning – Trip planning for the brainrot era [Travel]
4 points, 2 comments
A Tinder-style trip planning app for swiping through attractions and creating itineraries.
- Swipe-based attraction selection
- Drag-and-drop itinerary planner
- Note-taking for each place
- Automatic travel time estimation
- Itinerary auto-arrangement (limited effectiveness)
- Trip sharing with friends
- Open-source code available
Discussion Link, Original Link
Use Apple Container with Gemini CLI [Developer Tools]
4 points, 0 comments
Gemini CLI now supports Apple Containers for M1/M2/M3 Macs, extending its sandbox capabilities beyond Docker.
- Native Apple Container support for M-series Macs
- Extends Gemini CLI's sandbox functionality
- GitHub integration available for setup
Discussion Link, Original Link
Voice-Mode MCP – Conversational Coding for Claude Code, Gemini CLI [Developer Tools]
7 points, 0 comments
Voice-Mode MCP enables two-way voice conversations with AI coding assistants like Claude Code and Gemini CLI.
- Open-source MCP server for voice interaction
- Supports conversational coding workflows
- Compatible with Gemini CLI via configuration
- Demo available on YouTube
- GitHub repository for self-hosting