3/6/2026
Show HN: Aporium AI – a study app that refuses to give you answers
Aporium AI tutoring platform using Socratic method for student learning. LLM-based educational application with question-guided approach.
3/6/2026
Show HN: Rental Property Deal Analyzer – 20 metrics, deal scoring, AI analysis
Open-source rental property investment calculator using FastAPI and vanilla JS. AI analysis feature but domain is real estate, not AI/ML focus.
3/6/2026
Nested Training for Mutual Adaptation in Human-AI Teaming
Research on nested training approach for human-AI collaboration and mutual adaptation.
3/6/2026
Show HN: Tri·TFM Lens – 5-axis quality evaluation for ChatGPT/Gemini responses
Chrome extension evaluating ChatGPT and Gemini responses across 5 dimensions: emotion, fact, narrative, depth, and bias with quality scoring.
3/6/2026
Show HN: Claudine – A Kanban board for your Claude Code and Codex conversations
VS Code extension that creates a live kanban board from Claude Code and Codex agent session files with auto-detection and real-time sync.
3/6/2026
Show HN: I built the first scripting language for multiplayer game dev
Custom scripting language for multiplayer game development that abstracts networking details, designed to work well with AI code generation tools.
3/6/2026
Show HN: CloakPipe – Rust privacy proxy for LLM APIs with pseudonymization
Rust proxy for LLM APIs that detects sensitive entities, replaces them with consistent pseudonyms, and rehydrates responses. Preserves semantic structure for embeddings while protecting data privacy.
3/6/2026
Show HN: Mog, a programming language for AI agents
Mog: statically-typed programming language for AI agent plugins with capability system for enforcing permissions and I/O control.
3/6/2026
From vibes to data: measuring how LLMs attend to your prompt, layer by layer
Open-source mechanistic interpretability toolkit for analyzing LLM token attention and logit projections across layers with visualizations. Model-agnostic analysis.
3/6/2026
300 Founders, 3M LOC, 0 engineers. Here's our workflow
Two founders built 300k LOC codebase using 3-6 parallel AI coding agents. Non-traditional SWE backgrounds using agentic workflows.
3/6/2026
RepoSage – Understand any codebase in minutes using Claude or local Ollama
RepoSage uses Claude or Ollama to analyze codebases, generating structured English breakdowns and supporting interactive Q&A.
3/6/2026
Microsoft: North Korean agents using AI to trick western firms into hiring them
Microsoft report on North Korean agents using AI for social engineering in hiring. Security/threat intelligence focused.
3/6/2026
Show HN: A local, multi-agent, customizable stack built for researchers
Local Rust/Python multi-agent research stack with internet access, voice interface, and reproducible experiment workflows.
3/6/2026
Master Machine Learning with scikit-learn (free book)
Free practical guide to machine learning fundamentals and best practices using scikit-learn.
3/6/2026
Show HN: Not All Agents – convince a room of agents that you're one of them
Not All Agents is a terminal-based social deduction game where humans compete against AI agents powered by o4-mini with reasoning.
3/6/2026
Markdown, Llms.txt and AI Crawlers
Data analysis of AI crawler activity vs. citation traffic. Found crawlers fetched 1,241 pages per citation link sent.
3/6/2026
LLM-discussion: a local app for multi-model AI consensus (325 lines of Python)
LLM-discussion: 325-line Python app enabling multi-model consensus through local queries. Compares responses across multiple LLMs.
3/6/2026
Securing AI Agents: Why Traditional Authorization Isn't Enough
Analysis of authorization challenges specific to AI agents. Traditional access control inadequate for autonomous agent systems.
3/6/2026
Security Scanner for Agent Skills
Security scanner for validating AI agent skills. Addresses security vulnerabilities in agent capabilities.
3/6/2026
Show HN: Max – a federated data query layer for AI agents (and humans)
Max: federated data query layer for AI agents. Enables agents to query across multiple data sources without pulling into single database.
3/6/2026
Show HN: MyChatArchive – bring your full ChatGPT history into Claude via MCP
MyChatArchive imports full ChatGPT conversation history into Claude via MCP, enabling semantic search with local embeddings.
3/6/2026
Show HN: AdPipe – Generate 12 Meta ad copy variants from any product URL
AdPipe generates 12 Meta ad copy variants from product URLs using LLMs, covering multiple persuasion styles.
3/6/2026
Show HN: AI trading platform with 34% returns (3 months) – seeking acquisition
Autonomous trading system using LightGBM and JAX PPO for portfolio optimization. 34% returns over 3 months with real capital on $300/month infrastructure.
3/6/2026
The AI Agent Infrastructure Stack: Who's Building the Picks and Shovels
Analysis of AI agent infrastructure stack companies. Limited content available.
3/6/2026
Show HN: Autonomous AI platform that builds apps and tools automatically
SuperBuilder: open-source platform for AI agents to build apps, generate media, and orchestrate models. Includes plugin system and Docker setup.
3/6/2026
Show HN: Feedster, an RSS/feed reader focused on discovery and agent integration
Feedster: RSS reader with agent integration and discovery features. Combines feed aggregation with AI agent capabilities.
3/6/2026
The first AI agent worm is months away, if that
Analysis predicting arrival of AI agent worms within months. Security threats from autonomous agent self-propagation.
3/6/2026
Conductor – Scalable Workflow Orchestration Engine for Microservices
Conductor is a CLI tool for building multi-agent workflows with GitHub Copilot and Claude, supporting evaluator-optimizer loops and failure handling.
3/6/2026
OpenAI Symphony
Symphony spawns autonomous AI agents to handle work items from Linear board, completing tasks with proof of work and safe PR landing.
3/6/2026
Show HN: Argus – VSCode debugger for Claude Code sessions
Argus is a VSCode extension that monitors Claude Code sessions, tracking costs and detecting inefficiencies.
3/6/2026
The emerging role of SRAM-centric chips in AI inference
Technical analysis of SRAM-centric AI accelerators vs GPUs for inference. Compares Cerebras, Groq, d-Matrix architectures and memory tradeoffs.
3/6/2026
Imrobot – Reverse-CAPTCHA that verifies AI agents, not humans
Imrobot: reverse-CAPTCHA for verifying AI agents using deterministic string operation pipelines. Allows agents while blocking humans.
3/6/2026
Show HN: Dotclaude – Sync your Claude Code config across machines with Git
Dotclaude syncs Claude Code configuration across machines via Git with intelligent JSON merging and backup support.
3/6/2026
Show HN: Droidctx – generate Markdown docs from your prod infra for coding agent
Tool generating Markdown documentation from production infrastructure for AI agents to use as context during debugging and issue investigation.
3/6/2026
Show HN: RapidFire AI – parallel RAG experimentation with live run intervention
Developer tool for parallel RAG pipeline experimentation with concurrent configuration testing and live intervention capabilities.
3/6/2026
Tokf – CLI output compressor for AI coding agents
CLI tool that compresses output from AI coding agents to reduce token usage and costs.
3/6/2026
Agency Agents – complete AI agency with specialized experts
Collection of specialized AI agent personalities with different expertise areas, built iteratively for various tasks.
3/6/2026
Looking for Feedback on a Computer Agent
Computer agent accessible from desktop/phone with multi-model support, approval workflows, action recording, and sandbox execution environments.
3/6/2026
Show HN: Qurt – open-source desktop AI coworker (BYOK, agent mode)
Open-source desktop AI agent for Windows/macOS with bring-your-own-key support, local-first architecture, agent mode with web/terminal access.
3/6/2026
AI code review and PR summaries before you push
AI tool providing code review and pull request summaries before code is pushed to repository.
3/6/2026
Show HN: NERDs – Entity-centered long-term memory for LLM agents
NERDs: Wikipedia-style entity memory system for long-running LLM agents. Agents build networked entity documents to manage context window limits.
3/6/2026
Show HN: Traces: A new way to share and discover agent traces
Traces: Platform for capturing, sharing, and discovering AI agent execution traces with team collaboration and automation integration.
3/6/2026
OpenFelix – Open-source AI assistant for macOS (local MLX and any cloud model)
Open-source voice-first AI agent for macOS with local MLX support and cloud model backends, featuring tools and skills integration.
3/6/2026
Running production AI systems at scale (GoDaddy and AWS case study)
Case study of GoDaddy using AWS Bedrock AI tools to accelerate software development across their platform.
3/6/2026
Show HN: Claude skill to do your taxes
Claude Code skill that automates tax filing. Tested against TurboTax on 2024-2025 returns with matching results.
3/6/2026
Show HN: Anchor Engine – Deterministic Semantic Memory for LLMs Local (<3GB RAM)
Local-first semantic memory layer for LLMs under 3GB RAM. Provides persistent retrieval from user data without hallucinations.
3/6/2026
Show HN: Decidel A Hacker News client for iOS with smart summaries and filtering
iOS HN client with AI-powered summaries, semantic filtering, and offline reading. Brings own API key.
3/6/2026
Almanack of Agents: Dispatches from an AI Social Network
Discussion about AI agents operating in a social network environment. Limited details provided.
3/6/2026
Show HN: A document search engine CLI, built for AI Agents.
CLI tool for searching local documents via MCP server integration, designed for AI agents to access indexed documents from terminal.
3/6/2026
Show HN: ChronosBench – Terminal CPU/Metal GPU stress tester with live load bars
Terminal-based CPU/Metal GPU stress tester with live load bars, runnable in scripts and CI pipelines.