2/28/2026
Show HN: Cc-connect – Remote control Claude Code from your favorite chat app
Cc-connect: bridges local AI coding agents (Claude Code, Cursor, Gemini) to messaging apps without public IP.
2/28/2026
Watch your Claude Code hooks in real time
HookLab: real-time dashboard for monitoring Claude Code HTTP hooks with tool call inspection and filtering.
2/28/2026
Show HN: Cagent – Agent in a Cage
Cagent: Docker-based sandboxing tool constraining AI coding agents to isolated workspaces with network/filesystem restrictions.
2/28/2026
Show HN: GEKO (up to 80% compute savings on LLM fine-tuning)
GEKO: gradient-efficient fine-tuning optimizer that skips mastered samples and applies curriculum learning. Achieves up to 80% compute savings.
2/28/2026
How I'm Using Local Large Language Models
Personal experience article on local LLM deployment and usage. Practical guide for LLM applications without commercial LLM APIs.
2/28/2026
Claw21 – Blackjack for Agents
Blackjack game implementation for testing AI agents. Benchmark/playground for agent decision-making.
2/28/2026
Kbtz: A task tracker and TUI workspace for coding agents
Task tracker and TUI workspace designed for coding agents. Directly supports agentic AI workflows.
2/28/2026
Switch between different Claude Code profiles
Tool for managing multiple Claude Code profiles. Direct developer tooling for LLM-powered coding.
2/28/2026
Show HN: I dump all my private notes into an LLM and tell it to build me a site
Publishing platform that generates sites from unstructured notes using agentic LLM with minimal guardrails. Explores creative LLM applications.
2/28/2026
Show HN: Mdspec.dev – open-source spec management platform for technical teams
Open-source spec management platform optimized for agentic software development. Targets AI-assisted development workflows.
2/28/2026
Show HN: AIQuotaBar – See Claude/ChatGPT usage limits in your macOS menu bar
macOS menu bar utility tracking Claude and ChatGPT usage limits. Niche developer tool for LLM API management.
2/28/2026
Show HN: Voice-coded a Doom-scroll Hacker News (Twitter-style feed)
Hacker News infinite-scroll clone built entirely via voice coding using OpenAI Codex. Demonstrates LLM-assisted development workflow.
2/28/2026
Show HN: GitShow Repo Showroom – a landing page for any GitHub repo
GitShow generates showroom landing pages for GitHub repositories, aggregating contributor info, languages, commit activity, and health metrics.
2/28/2026
LLMs killed the privacy star, we can't rewind, we've gone too far
Research shows LLMs can deanonymize internet users more efficiently than humans, extending privacy risks beyond traditional de-anonymization methods.
2/28/2026
We measured 62% token reduction
LLM output verification and compliance tool achieving 62% token reduction with Python SDK and MCP integration.
2/28/2026
S2S – Physics-certified motion data for Physical AI (7 biomechanical laws)
S2S provides physics-certified motion data for training physical AI systems, validating sensor data against biomechanical laws for robots and prosthetics.
2/28/2026
Show HN: Using a mobile LLM app to safely operate a desktop computer
Computer-using AI agent that uses mobile LLM apps (ChatGPT/Claude) as intelligence source with external execution gateway for OS actions.
2/28/2026
Show HN: LLM-JSON-guard – Middleware to auto-repair broken AI outputs
Middleware to auto-repair and validate malformed JSON outputs from LLMs. Improves reliability of AI model outputs.
2/28/2026
Show HN: Mowgli – Figma for the agent era, with Claude Code and design export
AI-native design canvas with Claude Code integration and design export. Targets product scoping and ideation bottleneck in agent-driven development.
2/28/2026
Show HN: Enclv – A lightweight Docker containment environment for AI agents
Docker containment environment for running AI-generated code and autonomous scripts locally with network/file restrictions and credential isolation.
2/28/2026
Show HN: Shodh– AI memory that learns from use, no LLM calls, single Rust binary
Offline persistent memory system for AI agents that learns from use. Single Rust binary, no LLM calls required. Addresses context loss between sessions.
2/28/2026
Show HN: Fava Trails – Git-backed memory for AI agents using Jujutsu (JJ)
Git-backed memory system for AI agents using Jujutsu. Addresses memory poisoning through version-controlled agent state and history.
2/28/2026
Show HN: SQLite for Rivet Actors – one database per agent, tenant, or document
SQLite storage backend for Rivet Actors (Durable Objects alternative). Per-agent/tenant databases for AI agent state, message history, embeddings.
2/28/2026
Istota is a powerful (non-claw) AI agent that lives in Nextcloud
AI agent that integrates with Nextcloud as native user. Collaborates on files, manages calendars via Claude Code. Data stays on-premises.
2/28/2026
A Guide to Claude Code 2.0 and getting better at using coding agents
Guide on Claude Code 2.0 AI coding agent and techniques for effective prompt engineering and agent interaction.
2/28/2026
Mycelium is a Clojure framework designed to facilitate LLM coding
Clojure framework for building LLM-powered coding applications and agents.
2/28/2026
Garbage In, Garbage Out: The Degradation of Human Requirements in the LLM Era
Analysis of how LLM adoption changes workplace communication patterns and requirements specification quality.
2/28/2026
ClawShopping-Marketplace for AI Agents
Marketplace platform using Stripe where AI agents trade skills, build reputation through transparent signals, and interact with humans as observers.
2/28/2026
Show HN: Soma, a local-first AI OS with 178 cognitive modules and P2P learning
Local-first AI operating system with 178 cognitive modules, persistent memory, multi-model reasoning, and P2P learning network released as open source base.
2/28/2026
Show HN: IssueScout – Find open source issues worth contributing to
Tool that filters GitHub's open source issues by repository health score (based on 7 factors) to help beginners find actively maintained projects with genuinely beginner-friendly issues.
2/28/2026
Show HN: The most awesome AI programming application desktop has emerged
Desktop application for multi-agent AI collaboration that integrates with existing CLI tools, enabling parallel execution and orchestration without project migration.
2/28/2026
Show HN: Salacia – The First Runtime OS for Agentic Coding
Runtime OS for AI coding agents that maintains context across conversations, provides intent compilation, semantic drift testing, and auditable logs.
2/28/2026
Ask HN: Do fast-moving DevTools need localized documentation?
Discussion on localization strategy for developer tools and open-source projects with global users.
2/28/2026
Show HN: CocoSearch – semantic code search with syntax-aware chunking
Semantic code search tool with syntax-aware document chunking.
2/28/2026
Is GitHub Copilot still relevant in the enterprise?
Discussion comparing GitHub Copilot adoption against Claude Code, Cursor, Devin and other AI coding tools in enterprise settings.
2/28/2026
Prompt Repetition Improves Non-Reasoning LLMs
arXiv research showing that repeating prompts improves performance of non-reasoning LLMs, with empirical findings on prompt repetition techniques.
2/28/2026
AI assisted coding
Blog post on using LLMs as primary code editing tool, noting significant improvements in coding capabilities while emphasizing human authorship of prose.
2/28/2026
Show HN: Nano Banana 2 – 4K AI image generator with accurate text rendering
AI image generation platform using Google's Gemini 3.1 Flash model, addressing text rendering accuracy in generated images.
2/28/2026
Show HN: Rust-powered document chunker for RAG – 40x faster, O(1) memory
Rust-based document chunking library for RAG pipelines with Python bindings, 40x faster than LangChain with O(1) memory complexity.
2/28/2026
Sandboxed Git worktrees for AI coding agents
Workmux provides sandboxed Git environments for AI coding agents in containers/VMs, isolating agents from host credentials while maintaining RPC sync with dashboard.
2/28/2026
Show HN: MBTI personality test that AI agents take by themselves
Tool that runs MBTI personality tests on different AI models (Claude, GPT-4, smaller models) via agent-executed questionnaire with visual result breakdowns.
2/28/2026
Show HN: Memrail – PR-style governance for AI agent writes (OpenClaw)
Open-source governance layer for OpenClaw AI agent workflows treating memory writes as pull requests with dry-run, diffs, human approval, audit trails, and rollback.
2/28/2026
Seeking Advice on Improving OCR for Watermarked PDFs in My RAG Pipeline
Technical troubleshooting of OCR degradation in RAG pipelines when PDFs contain watermarks using PyMuPDF.
2/28/2026
Show HN: OpenPencil - Open-source vector design tool controlled by AI Agents
MCP server synchronizing configuration files across Claude, Cursor, Gemini, and other AI agents from single YAML source.
2/28/2026
Fluvel – a Python framework for building modern and declarative UIs with PySide6
Fluvel is a Python framework for building declarative UIs with PySide6.
2/28/2026
Show HN: Scite MCP – Full-text scientific papers in your AI workflow
Local-first tool using Chrome DevTools Protocol and Discord to control AI coding setup remotely without cloud relays.
2/28/2026
Show HN: Decided to play god this morning, so I built an agent civilisation
WERLD is an open-source artificial life simulation where neural network-based agents evolve language, behavior, and social structures from scratch.
2/28/2026
Show HN: K8s LoadBalancer services via Fly.io tunnels
Kubernetes operator that provisions frp tunnels via Fly.io Machines to fulfill LoadBalancer services, designed for homelabs.
2/28/2026
Show HN: CarbonLint – Open-source real-time energy&carbon profiler for software
CarbonLint is an open-source Rust/Tauri tool that profiles software energy consumption and carbon emissions in real-time for developers.
2/28/2026
I don't pay for ChatGPT, Perplexity, Gemini, or Claude – I stick to my self-h
Article about self-hosting open-source LLMs locally via Ollama instead of relying on cloud-based AI services, with privacy and cost benefits.