If you've been hunting for a Markdown editor that doesn't eat your RAM for breakfast while still giving you local AI muscle, VelocityNote might be worth bookmarking. This cross-platform notebook app just showed up on Hacker News with a pitch that's refreshingly straightforward: under 100MB, fast startup, and built-in support for running LLMs locally through llama.cpp.

What You're Getting

VelocityNote positions itself as a compact alternative to the bloated note-taking giants. The desktop binary stays comfortably under 100 megabytes—a rarity in an era where Electron apps routinely balloon past 300MB before you even open a document. Startup time is allegedly snappy, and the developer claims it remains responsive when memory pressure kicks in. For developers who keep half a dozen tools running simultaneously, that kind of efficiency isn't trivial.

The Local AI Angle

The hook here is the built-in AI system with pluggable model support. Rather than shipping you off to OpenAI or Anthropic's APIs, VelocityNote can run models locally using llama.cpp. That means your notes stay on your machine—no telemetry, no subscription fees, no 'improving our services' fine print. The architecture appears designed around flexibility: swap in different models depending on what you're trying to accomplish.

Why This Matters for Builders

Markdown notebooks with local AI are having a moment, and for good reason. When you can summarize meeting notes, brainstorm ideas, or refactor code snippets without touching the cloud, you sidestep latency, privacy concerns, and API costs in one shot. A sub-100MB app that handles this cleanly fills a gap between heavyweight IDEs and basic text editors.

Early Days

Fair warning: the HN post is sitting at a score of 2 as I write this. The project is fresh, community feedback is minimal, and real-world stress testing hasn't happened yet. But if the claims hold up—fast startup, memory efficiency, genuinely local AI—this could be the kind of tool that earns its reputation through utility rather than hype.

Key Takeaways

  • Cross-platform Markdown editor with desktop app under 100MB
  • Built-in llama.cpp integration for running LLMs locally
  • Designed to stay responsive under memory pressure
  • No cloud dependency—everything stays on your machine
  • Early-stage project, limited community validation so far

The Bottom Line

If you're a developer tired of cloud-dependent AI features eating into your privacy budget, VelocityNote's approach is worth watching—provided the performance claims survive real-world use. A sub-100MB editor with local LLM support could genuinely change workflows for those who value speed and data sovereignty over flashy integrations.