Interactive demos for LiteParse — a fast, local, model-free document parser. Includes a parser comparison (LiteParse vs PyPDF vs PyMuPDF), visual citation search with bounding boxes, and a Claude Code research skill.
Anthropic’s Claude Mythos Preview can autonomously find and exploit zero-day vulnerabilities in every major OS and browser. Here’s what that means — and what Project Glasswing is doing about it.
Recordly is an open-source desktop screen recorder/editor with timeline tools, cursor polish, webcam overlays, and MP4/GIF export.
Factual summary of Xiaomi MiMo-V2-Pro launch claims, benchmark positioning, and API pricing versus selected frontier models.
Build log for Research Podcast Studio: SSE streaming UX, revision prompts, inline status, debugging token-thinking and reconnect loops, and end-to-end pipeline validation.
Follow-up update with an improved dual-voice audio sample generated from the CBC story workflow.
Interesting to note perhaps, that I did all of this – including this post – from my cellphone on the bus ride home. I wanted to see whether I could build a practical “research-to-podcast” workflow inside OpenClaw without over-engineering it. This post documents what I did, what worked, what broke, and what I would improve…
I ran a controlled trial of desloppify inside my OpenClaw workflow to test whether it provides useful, non-noisy quality signals on a real repository. How I ran this via OpenClaw Selected a real repo for trial: /home/adrian/code/apipairy. Created an isolated branch for safety: test/desloppify-20260310. Installed desloppify in a local virtual environment (no global system mutation)….
Cozy 3D Room is a compact interactive demo built with Three.js and WebGL. According to the About section, the project is implemented as a single self-contained HTML file with no build tools, no server dependencies, and no external assets beyond the Three.js CDN. The demo includes clickable object interactions (ceiling fan, lamp, drawer, and framed…
Overview This is a factual record of a structured multi-model review run completed on 2026-03-07 for the hensler-photography project and live site surfaces. Method used: independent model critiques (DeepSeek, Mistral, Qwen), followed by Sonnet 4.6 arbitration and final synthesis. Prompt Pattern Used 1) Independent review prompt You are performing a critical architecture + product review…