3/18/2026
Qianfan-OCR – 4B open-source VLM replacing multi-stage OCR pipelines
Qianfan-OCR: 4B-parameter open-source vision-language model for document understanding and image-to-Markdown conversion.
3/18/2026
Nvidia NemoClaw
Nvidia NemoClaw open-source stack for running OpenClaw autonomous assistants safely with sandbox orchestration.
3/18/2026
Show HN: Agent Trust – Cryptographic identity and reputation for AI agents
Agent Trust: System providing cryptographic identity and reputation mechanisms for AI agents. Minimal details in snippet.
3/18/2026
Show HN: Habits, an open-source builder and runtime for automations and apps
Habits: Open-source graph-based system for building automations, AI agents, and apps from unified model. Exports to Docker, binaries, desktop and mobile apps.
3/18/2026
Why AI agents can reduce productivity
Discussion on potential productivity drawbacks of AI agents. Limited substantive technical content.
3/18/2026
Show HN: Lukan – An open-source agentic workstation in a single Rust binary
Lukan: Open-source MIT-licensed agentic workstation as single Rust binary with zero dependencies. Includes secure remote access, file management, and integrated AI agents.
3/18/2026
Show HN: A 30-minute course to get up to speed on how AI agents work
30-minute interactive course teaching AI agent fundamentals through 60 lines of Python, browser-based.
3/18/2026
Show HN: PlanckClaw an AI agent in 6832 bytes of x86-64 assembly
PlanckClaw: minimal AI agent implementation in 6.8KB of x86-64 assembly for tool-calling workflows.
3/18/2026
Show HN: Permit MCP Gateway – Fine-Grained Authorization and IGA for MCP
Permit MCP Gateway: authorization proxy for Model Context Protocol servers with fine-grained access control.
3/18/2026
Show HN: On-device meeting transcription for your Mac
On-device meeting transcription app for macOS using Apple Core Audio APIs and local processing.
3/18/2026
570k Lines of LLM Code Compiled Fine. It Was 20,171x Slower Than SQLite
Benchmark showing LLM-written Rust SQLite reimplementation compiled correctly but ran 20,171x slower than original.
3/18/2026
Ask HN: How is your company managing internal AI agents?
Discussion thread on operational management of custom internal AI agents in companies.
3/18/2026
Is there an AI garage startup path?
Analysis of startup viability in AI foundation models given computational and capital barriers.
3/18/2026
Show HN: Atria – terminal UI for managing multiple coding agents
Terminal UI multiplexer for managing multiple coding agent sessions with status visibility and quick controls.
3/18/2026
Build Memory-Aware Agents
Article about building agents with memory capabilities. Title only, minimal content provided.
3/18/2026
Mining your team's PR reviews into automated code review rules
Article about extracting code review patterns from PR history to create automated review rules.
3/18/2026
Build AI Agents for Elevation, Not Replacement
Opinion piece on designing AI agents for human capability enhancement rather than replacement.
3/18/2026
Show HN: SHTMLs – HTML pastebin where the AI uploads its own output
SHTMLs: HTML pastebin for sharing Claude-generated outputs with programmatic API for AI agents.
3/18/2026
Show HN: Deploybase CLI – Search GPU and LLM pricing from your terminal
CLI tool for searching and comparing GPU and LLM pricing across 100+ cloud providers with live data.
3/18/2026
Show HN: Vibe – a language for humans and AI to reason about programs together
Vibe language for Scheme-to-LLVM compilation with AI assistance for co-reasoning about programs.
3/18/2026
Show HN: Ranvier – Prefix-aware routing for LLM inference
Ranvier enables prefix-aware routing optimization for LLM inference to improve performance.
3/18/2026
Detecting When Your AI Agent Dies
Title only, no content provided. Likely about monitoring AI agent failures or health checks.
3/18/2026
Why Agent UIs Lose Messages on Refresh
Technical explanation of UI state management issues when AI agent chat interfaces refresh and lose message history.
3/18/2026
Running 9 AI agents and realizing we needed a control plane
Experience report on operating multiple AI agents and identifying the need for a centralized control plane.
3/18/2026
Show HN: Ibkr-CLI – A local-first Interactive Brokers CLI built for AI agents
Ibkr-cli is a CLI tool for Interactive Brokers designed for AI agents to execute trading actions programmatically without the legacy desktop interface.
3/18/2026
Veto: Permission policy engine and LLM firewall for AI coding agents
Veto is a permission policy engine and LLM firewall designed to control actions in AI coding agents.
3/18/2026
Heaps do lie: debugging a memory leak in vLLM
Technical debugging guide for memory leak in vLLM, a widely-used LLM inference optimization library.
3/18/2026
Show HN: Applemail – Cursor skill for full macOS Mail.app control
Applemail is an AI agent skill enabling Cursor to control macOS Mail.app for email management tasks through natural language.
3/18/2026
Show HN: Elisym – Open protocol for AI agents to discover and pay each other
Elisym is an open protocol enabling AI agents to discover each other, exchange work, and settle payments autonomously using Nostr relays and pluggable payment backends.
3/18/2026
Show HN: Hipocampus – Persistent memory harness for AI agents
Hipocampus is a persistent memory harness tool for AI agents, enabling agents to maintain state across interactions.
3/18/2026
Show HN: Faramesh – open-source runtime enforcement for AI agents
Faramesh is an open-source runtime enforcement tool for AI agents that intercepts tool calls before execution to evaluate decisions against policies, addressing production safety gaps.
3/18/2026
Show HN: Site got cited by Claude and ChatGPT in under a week
Website citations in Claude and ChatGPT generated referral traffic within days of publication using llms.txt and schema.
3/18/2026
Show HN: OpenGranola – meeting copilot that searches your notes in real time
OpenGranola is an open-source macOS meeting copilot that transcribes conversations locally and searches notes in real-time.
3/18/2026
Study: Top AI coding tools make mistakes one in four times
Research study shows top AI coding assistants produce mistakes approximately 25% of the time.
3/18/2026
Show HN: Mimir – open-source code intelligence for AI agents (Go, MCP, SQLite)
Mimir is open-source code intelligence that builds typed knowledge graphs for AI agents with AST parsing across 8 languages via MCP.
3/18/2026
Show HN: CLI to score AI prompts after a prod failure
CLI tool for scoring and validating AI prompts in production to prevent failures from instruction ambiguities.
3/18/2026
Show HN: A single pane of glass for Claude Code
IDE alternative integrating Claude Opus 4.6 with unified interface for managing markdown, configs, and AI coding workflows.
3/18/2026
MiniMax M2.7 (200K context, $0.30/1.20) released
MiniMax M2.7 is an LLM with 200K context window and agentic capabilities for autonomous task execution, debugging, and document generation.
3/18/2026
Show HN: Open-source status pages for Agents (Rails 8 and SQLite)
ups.dev is an open-source status page alternative built with Rails 8 and SQLite for monitoring and incident management.
3/18/2026
Show HN: Cloak – .env on disk has fakes, your editor shows them (CLI and VSCode)
Cloak protects local .env files with real credentials from being exposed to AI agents with filesystem access by showing fake values in editors while keeping real secrets on disk.
3/18/2026
Show HN: We built AI agents that reduce mortgage processing from 18 days to 3–5
SimplAI uses AI agents to automate mortgage processing workflows, reducing 15-22 day timelines to 3-5 days by handling document processing and data verification.
3/18/2026
I built an agent that tells you when a GitHub bug is killing your Stripe revenue
Agent that monitors GitHub bugs and correlates them with Stripe revenue impact, demonstrating practical LLM agent application.
3/18/2026
Chinese users are using an imperial China court system to manage AI agents
Article discussing how Chinese users employ historical imperial court system frameworks to manage and govern AI agent behavior.
3/18/2026
Show HN: Zora, AI agent with compaction-proof memory and a runtime safety layer
AI agent framework with compaction-proof memory and runtime safety layer, addressing constraint loss during context compression.
3/18/2026
Mamba-3
Research paper on Mamba-3, a state space model optimized for inference efficiency as primary design goal, with 40+ production models.
3/18/2026
Ask HN: What's the costliest mistake you've made with LLM agents in production?
Discussion thread about costly production mistakes with LLM agents, including cost overruns and prompt injection vulnerabilities.
3/18/2026
AI agent can't tell if what it reads is TRUE/FALSE. I built an API to fix that
API providing fact verification and world knowledge signals for AI agents across 18 vertical categories, updated in real-time.
3/18/2026
Show HN: Reprompt – Score your AI coding prompts with NLP papers
Tool that scores AI coding prompts using NLP research-backed features, running locally without API calls for privacy.
3/18/2026
Claude-to-IM Skill: Bridge Claude Code / Codex to Telegram, Discord
Open-source bridge integrating Claude Code and Codex with messaging platforms (Telegram, Discord, Feishu, QQ) with safe permission prompts.
3/18/2026
Show HN: Precise AI Motion Control for Kling 3.0
AI Motion Control tool for Kling 3.0 video model enabling deterministic camera movements like orbits and crane shots.