First formally verified polygon intersection—Lean 4 proofs guarantee correctness for all inputs.
Wire-protocol parsing gates agent actions before they hit production—no LLM gateway does this.
Container isolation and tmux sessions for async multi-agent workflows.
Pretty shader that the author admits won't make you more productive.
AI wrapper on parametric search when DigiKey and Octopart already exist.
A threat-intelligence approach to VC meetings. Six prompts, ~25 min per investor. Built with Claude Code.
Threat-intelligence methodology applied to VC prep with strict no-guessing rules.
Finally replaces the 25-year-old Enron corpus with deterministic org simulation.
Personality quiz funneling users to an App Store app, nothing more.
Fine-tuned 0.6B model keeps agent knowledge graphs clean using dynamic text rules.
Sessions persist across Claude accounts when one hits rate limits.
A police department for your Claude Code agents — a logging-only hook + CLI that audits the main agent and every subagent and reports rule offenses (permission bypass, out-of-scope & credential access, self-permissioning, disallowed tools, redundant, off-task) with quoted evidence. Catch-and-report, never blocks.
Hash-chained audit log catches agent violations at zero token cost.
Cited evidence for scam verdicts when most checkers are black boxes.
Queue psychology mechanic where walking feels faster than standing still is clever game design.
Personal gift project, not a tool others can use or replicate.
Auto-discovers Claude PRs via gh CLI and jumps into originating sessions.
Curated prompt library for co-founders, no technical innovation.
Write-once-read-once mechanic exists in Snapchat and Telegram already.
Yet another AI playlist generator, but at least it's ad-free.
File-level permissions and live events fix Git's agent-team gaps.
Full LLM pipeline in one window when LM Studio only does inference.
C2PA inspector and true PDF redaction stand out in a sea of basic EXIF removers.
Interactive cable tracing between racks shows power paths visually in React.
Federates LandXML, IFC, and GeoJSON in one scene without format conversions or data loss.
Card-based interface beats spreadsheet tables for mobile database browsing.
Explore and simulate the 2026 FIFA World Cup — typed tournament data, an LLM simulation kernel, and in-browser TTS commentary.
Four-agent match simulation with in-browser TTS commentary via WebGPU.
DevTools console injection beats MSW's codebase integration requirement.
Email threading as state memory eliminates vector stores and message queues.
Local MCP server that gives AI coding agents persistent, searchable memory of your codebase
76% token reduction on real projects with one-command setup.
Local tokenizer execution beats sending text to external counters.
Zero-trust, capability-based Rust microkernel targeting formal verification. Tri-arch (x86_64 / AArch64 / RISC-V). Sovereign and generative: no telemetry, user owns keys and data. Early-stage — see STATUS.md. Inspired by seL4, Hubris, and Redox.
Ambitious Rust microkernel OS, but seL4 and Redox already own the verification-ready space.
Web-scale and security-hardened API key server for users, services, machine to machine, and AI agents. Token derivation brings fine-grained capability tokens to avoid common API key pitfalls. Apache2 open source for indie deployments, commercial for scalable and HA.
Token derivation for fine-grained capabilities beats static API keys from Auth0 or Vault.
Glues Google TTS and Liquidsoap together for dynamic personal radio stations.
Parallel Go probes catch intermittent latency smokeping misses.
Ahrefs API wrapped in a retro leaderboard — TrustMRR but for domain rating.
AI code reviews grounded in 12 classic engineering books — decay risk diagnostics with book citations, severity labels, and 6 analysis modes including full-sweep auto-fix
Grounds AI code reviews in 12 engineering classics with book citations.
Yet another website chatbot, but with a mascot skin and voice.
Make.com integration is convenient, but Personal Capital and TradingView already handle portfolio alerts.
Finds the code your AI wrote twice. Fast, offline duplicate-code detector: scan, history chart, CI gate. No AI required.
History chart shows exactly when your code duplication spiked — SonarQube doesn't do this.
Control plane with token budgets and MCP rules goes beyond passive observability.
Git worktree isolation lets agents build in parallel without stepping on each other.
Real-time LLM token and cost monitor with TLS-intercepting proxy or HTTP relay; cross-platform with macOS status bar app and browser dashboard
Wire-protocol proxy means zero code changes to existing LLM clients.
Spawn and manage AI agents that talk to each other. systemd-managed multi-agent runtime on your own VM.
Using Linux users for agent isolation is clever, but Telegram limits appeal.
Free open-source Google Maps scraper — Chrome extension to extract business leads (name, address, phone, website, email) from Google Maps, Yandex Maps & 2GIS and export to CSV/Excel. No API key, no account.
Free Maps scraper when Apify and ScraperAPI already dominate this space.
Line under cursor stays raw while everything else renders in place.
Yet another AI idea validator when ChatGPT already does this for free.
Finally solves sharing agent artifacts without pasting raw Markdown files.
Visually striking map with 4K+ predictions and zero signup friction.
AIOps Infra for deploy and manage self-hosted local AI in your own cloud. Vibe coding and AI agents compatible.
Yet another AI infra platform when Modal and RunPod already dominate.
A macOS menubar app that sends screenshots and short recordings directly into a running Claude Code iTerm2 terminal session. Also supports sending screenshots from an iPad/iPhone on the same WiFi network.
Injects screenshots into Claude Code sessions via hotkey or iOS Shortcut bridge.
743 projects
