GitHub 2026-08-14 Daily Trending Projects

A self-contained editorial diagram library for docs and agent skills — 27 visual types, brand auto-matching, and no-JS static output for fast, consistent, and accessible diagrams.

Semantica is a graph-native infrastructure for regulated environments, delivering auditable decision objects, deterministic reasoning, and ontology governance to turn enterprise data into explainable, traceable context graphs for compliance and interpretability.

Anthropic/skills is a collection of skill specifications and examples for Claude, offering reusable templates and document-skill references suitable as a foundation for custom skill development; however, license clarity and community activity should be assessed before production adoption.

Needle 2 is an ultra‑compact model focused on tool calling and constrained decoding for structured extraction; a 14MB single file that enables offline edge inference and low‑resource deployment, though license and maintenance risks should be assessed.

FluidVoice provides privacy-first, on-device voice-to-text and command control for macOS, combining multi-model low-latency transcription with optional local AI post-processing—ideal for users demanding real-time interaction and no cloud data.

Unsloth delivers a memory-friendly fine-tuning and RL toolchain for research and engineering, supporting multiple models and free notebooks—great for low-cost prototyping and experiments, but production readiness and performance reproduction require careful evaluation.

Macro aims to be an all-in-one workspace for small teams, connecting email, messages, docs, tasks and CRM with shared team memory to reduce tool switching and improve collaboration; however, the public repository currently lacks active maintenance, releases and license information, so risk and compliance checks are required before adoption.

Holehe is an OSINT tool that maps emails to registered accounts by probing password-recovery and login endpoints across 120+ services. It offers a CLI and asynchronous Python modules for integration, and is best applied in controlled, legally compliant workflows such as threat intelligence, red teaming, and privacy audits.

SpiderFoot is a Python-based OSINT automation platform using 200+ modules and correlation rules to discover, correlate, and visualize externally exposed assets and threats for reconnaissance and defensive use.

Switchyard is an experimental Rust proxy/library offering multi-backend LLM routing, OpenAI/Anthropic API translation and Prometheus metrics for local model integration and A/B testing; currently pre-alpha with low repository activity.

holaOS delivers a local-first multi-agent workspace with editable shared memory and pluggable model access, emphasizing privacy and control—suited for teams needing multi-model coordination, self-hosted deployment, and enterprise integrations.

A collection of Agent Skills for Obsidian that provides standardized editing and automation skills for reuse across agents like Claude Code, Codex CLI, and OpenCode; useful for integration but requires careful review of license and maintenance transparency.

Manim, originated by 3Blue1Brown, is a programmatic math-animation engine that produces reproducible, high-precision instructional visuals; ideal for educators and creators with Python and CLI experience for math visualization and outreach.

Persona-centric AI agent collection with production-ready Claude Code templates for rapid specialist assistants; license and maintenance transparency are lacking.

LTX-2 is a DiT-based production audio-video model emphasizing synchronized audio and high-fidelity outputs for compute-heavy teams.

Modly is a desktop tool for local GPU inference that converts photos into exportable 3D meshes via extensible AI workflows, aimed at developers and artists needing self-hosted generation.

RAGFlow is an open-source, enterprise-focused RAG engine that fuses retrieval-augmented generation with Agent capabilities to offer configurable document ingestion, chunking, embedding and re-ranking pipelines for building high-fidelity, traceable knowledge layers and QA systems; however, pay attention to license metadata, code activity, and runtime resource needs.