Discovery
Back to browse
GitHubTool

vdirsyncer - sync calendars and contacts

Python tool that mirrors CalDAV/CardDAV servers to local files (and back). The plumbing layer behind half the FOSS calendar setups out there.

View source ↗
#caldav#carddav#calendar#contacts#python

This entry doesn't have a long-form writeup yet. Follow the source link above for the full context.

Related entries

GitHubToolFeatured

Rapid-MLX - 2-4x faster local LLM inference on Apple Silicon

MLX-native inference engine with OpenAI-compatible API. The novel piece: DeltaNet state snapshots bring prompt caching to non-trimmable architectures (Qwen3.5 hybrids), restoring RNN state in ~0.1ms. 2-5x faster TTFT, native Metal kernels, continuous batching.

Why I saved this - DeltaNet state snapshots are described as the first prompt-cache technique for non-trimmable MLX architectures - Ollama and llama.cpp can't match this on hybrid RNN-attention models today.
#developer-tools#local-llm#apple-silicon#python#self-hosted
GitHubToolFeatured

PostTrainBench - can a CLI agent post-train a base LLM in 10 hours?

Benchmark measuring whether Claude Code, Codex CLI, Gemini CLI, and OpenCode can autonomously improve 4 small base models (Qwen3-1.7B/4B, SmolLM3-3B, Gemma-3-4B) on 7 evals (AIME, BFCL, GPQA, GSM8K, HealthBench, HumanEval, Arena Hard) within a single H100 GPU and 10 hours. Includes agent-as-judge anti-reward-hacking and baseline-replacement penalties for tampering.

Why I saved this - Current leader: Opus 4.6 via Claude Code at 23.2 average. The reward-hacking safeguards (eval tampering and model-substitution detection, baseline-replacement penalty) are the part most agent benchmarks skip.
#evals#claude-code#codex#gemini-cli#opencode
GitHubToolFeatured

mcptube - Karpathy-style LLM wiki for YouTube

MCP server that turns YouTube videos into a persistent, merging wiki rather than ephemeral vector chunks. Scene-change frame extraction + vision analysis captures slides, code, and diagrams that transcripts miss. 25+ MCP tools, FTS5+LLM hybrid retrieval, version history with source attribution per claim.

Why I saved this - The wiki-merge design is the differentiator vs RAG-over-YouTube clones - one MCP article with citations, not ten near-duplicate chunks. Scene-change extraction is what makes visual-heavy talks usable.
#mcp#claude-code#knowledge-graph#agent-memory#python