26d ago
Cosmos-Predict2.5-2B Inference
Minimal content on video prediction model. No technical details.
26d ago
Testing suggests Google's AI Overviews tells lies per hour
Analysis of Google AI Overviews search accuracy limitations. Not developer-focused.
26d ago
Show HN: Frontend-VisualQA — give coding agents eyes to verify their own UI work
CLI + MCP server tool enabling coding agents visual verification of UI layouts via browser-based testing.
26d ago
Frequent ChatGPT users are accurate detectors of AI-generated text (2025)
Study on humans detecting AI-generated text. No technical depth or code provided.
26d ago
Show HN: Coding Planets – turn GitHub contributions into sound and space
Visualization tool converting GitHub contribution history into procedural planets and generative music using three.js and tone.js.
26d ago
Show HN: RefearnApp – Open-source alternative to Rewardful/Tolt
Open-source, self-hostable affiliate tracking and commission management tool for referral programs.
26d ago
Prompt Caching from First Principles, blog with an AI co-author
Technical guide on prompt caching optimization techniques with AI co-authoring.
26d ago
Show HN: I built a database for AI agents
DuckDB-based database system for SQL-capable AI agents with benchmarks across 11 LLMs.
26d ago
Ask HN: How do you promote apps which are vibe coded but has real life usecase?
Discussion on marketing developer tools built with rapid development practices.
26d ago
We no longer write any of the prompts in our codebase
Tool that automatically discovers optimal system prompts for LLM tasks by analyzing desired output examples, eliminating manual prompt engineering.
26d ago
Hazmat: OS-level containment for AI coding agents on macOS
OS-level containment system for AI coding agents on macOS, addressing security risks when running untrusted agent code with filesystem/system access.
26d ago
Show HN: Build queryable packs for AI agents from videos, podcasts, and files
Developer tool that creates queryable knowledge bases from videos/podcasts for AI agents.
26d ago
Dexter Hadley on building governed AI for healthcare – Achieve Podcast [video]
Podcast discussion on governed AI systems in healthcare domain.
26d ago
AI Agent Sandboxes Got Security Wrong
Discussion of security vulnerabilities in AI agent sandbox implementations.
26d ago
Show HN: Hazmat – I made unrestricted Claude Code safe on macOS
macOS containment system using kernel sandboxing and firewalls to safely run unrestricted Claude Code agents autonomously.
26d ago
Show HN: I built an AI that forgets things when people leave the room
AI assistant prototype with session-aware memory that forgets context when users leave.
26d ago
AI Agent Guardrails: Pre-LLM and Post-LLM Best Practices
Best practices guide for AI agent guardrails, covering pre/post-LLM safety patterns.
26d ago
Same LLM but different output: we built a CI specialist
Mendral is a CI specialist coding agent built on Claude. Demonstrates how identical LLMs produce different outputs through system prompts, tools, and context optimization.
26d ago
Good Taste the Only Real Moat Left
Opinion on competitive advantages in AI era focusing on design taste.
26d ago
Ask HN: Why isn't Anthropic eating their own dogfood? A Max subscriber's view
Critique of Anthropic's own AI implementation practices versus enterprise recommendations.
26d ago
Are Local LLMs Good Enough for Vibe Coding? Gemma4-26B-A4B vs. Qwen3.5-35B-A3B
Compares local LLMs (Gemma4-26B, Qwen3.5-35B) for agentic coding tasks using OpenCode and Pi-Coding-Agent with custom tool usage scenarios.
26d ago
I Built a Wiki LLM
A wiki-based LLM system built on CIS security controls documentation, enabling semantic search and knowledge retrieval over structured security frameworks.
26d ago
Show HN: Developed a solution to automate ASO keyword research
RespectASO automates app store optimization keyword research with features for keyword difficulty and opportunity scoring.
26d ago
Show HN: Video Commander – A desktop IDE for video engineers (FFmpeg, VMAF)
Video Commander is a desktop IDE for video engineering combining FFmpeg, VMAF and other tools. Built with Tauri and React.
26d ago
Show HN: Falcon Builder – the modular OS for designing and deploying AI agents
Open source modular OS framework for designing and deploying AI agents. Show HN submission with practical agent infrastructure.
26d ago
AI alignment: the signal is the goal
Brief mention of AI alignment research. Insufficient detail provided.
26d ago
Show HN: ErrataBench - A Proofreading Benchmark for LLMs
ErrataBench is a benchmark measuring LLM proofreading performance across 51 model variants using an agent loop, with detailed runtime and cost metrics.
26d ago
I don't want an autonomous AI agent. I want a collaborator
Discussion of LLM collaboration patterns in developer tools like Cursor and Claude. Explores user experience challenges with autonomous AI agents.
26d ago
Building EU payment infra for agents
Infrastructure platform for payment processing integrated with AI agents in EU.
26d ago
Show HN: Testreel – Programmatic product demo videos from JSON and Playwright
Testreel: npm package for programmatic demo video generation from JSON/YAML/Playwright. Enables LLM agents to create product demos with cursor overlay and customizable backgrounds.
26d ago
Show HN: Opensidian: Local-first notes in the browser with POSIX shell and sync
Local-first note-taking app in browser with shell and sync. Developer tool but not AI-focused.
26d ago
Ace-Step-1.5 playground- SOTA open source music model
State-of-art open source music generation model playground. ML research but outside core interests.
26d ago
Claude is having another moment, again
Vague commentary on Claude's popularity/performance. No technical substance.
26d ago
How to Build an AI Agent for Slack with Chat SDK and AI SDK
Tutorial on building AI agent for Slack using Chat SDK and AI SDK. Developer guide for LLM integration.
26d ago
Claude Login Down?
Service status inquiry for Claude API. Not technical content.
26d ago
Show HN: OneManCompany The first AI company with real corporate org structure
Multi-agent system organized as functional company with independent AI agents in HR, engineering, design roles. Novel agent architecture approach.
26d ago
Show HN: Output.ai - OSS framework we extracted from 500+ production AI agents
Open source framework extracted from 500+ production AI agents. Production-tested patterns and tools for building agents at scale.
26d ago
Resurrecting a 1992 MUD with AI: No Source Code, Just 30-Year-Old Artifacts
Case study: Using AI to reconstruct and resurrect a 1992 MUD game from artifacts and old documentation. Technical restoration project.
26d ago
Show HN: SQLite Memory – Markdown based AI agent memory with offline-first sync
SQLite extension providing persistent searchable memory for AI agents with vector search, markdown support, and offline-first sync. Open source.
26d ago
OpenClaude – Forked the leaked Claude Code source and made it work with ANY LLM
Open-source CLI coding agent supporting multiple LLM providers (OpenAI, Gemini, Ollama, local models) with MCP and streaming.
26d ago
Show HN: Jsont – CLI to explore and query JSON/JSONL files, built for AI agents
CLI tool for querying JSON/JSONL files, specifically built for AI agent workflows. Open source developer utility.
26d ago
Lemonade 10.1 Released for Improvements for Local LLMs on AMD GPUs and NPUs
Lemonade 10.1 release with optimizations for running local LLMs on AMD GPUs and NPUs.
26d ago
Show HN: A tool to turn random screenshots into structured tutorials
React/TypeScript tool for converting screenshots into annotated step-by-step visual guides with sharing.
26d ago
Give an LLM an API and It'll Thrive. Give It a Touchscreen and It Struggles
Research comparing LLM performance differences between API-driven and GUI-based (touchscreen) interaction modes.
26d ago
Decentralized Training Can Help Solve AI's Energy Woes
Discusses decentralized training approaches to reduce AI model training energy consumption.
26d ago
Only 28% of AI infrastructure projects pay off, survey finds
Survey reporting that only 28% of AI infrastructure projects achieve successful outcomes.
26d ago
Indxr v0.4.0 – Teach your agents to learn from their mistakes
Fast codebase indexer providing persistent wiki for AI agents to learn from mistakes. Open source tool for agent knowledge management.
26d ago
The UI of Agentic SaaS
Discussion of user interface patterns and design considerations for agentic SaaS applications.
26d ago
Show HN: JavaScript runtime instrumentation via Chrome DevTools Protocol
Chrome DevTools Protocol-based JavaScript runtime instrumentation tool for debugging and execution interaction.
26d ago
Show HN: SolidUptime – Uptime monitoring with incident grouping (free, no CC)
Uptime monitoring service with incident grouping. Infrastructure tool unrelated to AI.