6 Must‑Watch Open‑Source Projects Transforming AI, Dev Tools & Media

1. RenderCV – YAML‑Based Professional Resume Generator

RenderCV solves the tedious resume‑formatting problem by letting users fill a simple YAML file with personal information, education, and work experience, then automatically producing a polished PDF using a LaTeX engine under the hood. No LaTeX knowledge is required.

The tool separates content from presentation, supports multiple built‑in themes for programmers and academics, and allows quick style changes via a single configuration line. Because the resume is plain text, it works well with version control for tracking changes.

Key metrics: 13.3k GitHub stars, 135k PyPI downloads, current version v2.6, and 99% test coverage. Open source address : https://github.com/rendercv/rendercv

2. Vibe‑Kanban – Lightweight AI‑Powered Task Board

Vibe‑Kanban reimagines task management for the AI era. It combines an intuitive drag‑and‑drop UI with deep AI integration that automatically parses project status and transforms vague ideas into concrete, actionable cards, eliminating manual card creation.

Designed for small teams and solo developers, it supports fast search, ESC‑key shortcuts, and a modern minimalist interface. Unlike heavyweight tools such as Jira, Vibe‑Kanban acts like an intelligent assistant that understands developers’ workflow, especially those using AI coding assistants like Cursor.

Open source address : https://github.com/BloopAI/vibe-kanban

3. Claude‑Code‑Templates – Efficient Prompt Library for Claude

Claude‑Code‑Templates is a curated collection of prompts that streamline AI‑assisted programming with Claude. High‑quality prompts reduce the effort needed to obtain usable code from the model.

The repository (14.1k stars) offers 102 AI agents, 159 custom commands, 51 settings, 29 hooks, and 25 external service integrations, covering scenarios such as code audit, architecture refactoring, bug locating, API development, data processing, game development, and security auditing.

Installation is flexible: install globally via npm or invoke directly with npx. Users can mix and match components to build personalized development stacks, gaining guidance comparable to an experienced senior engineer.

Open source address : https://github.com/davila7/claude-code-templates

4. Skills – Anthropic’s Official AI Skill Repository

Anthropic’s Skills project provides a standardized set of “hands‑and‑feet” for AI models, enabling Claude and similar agents to interact with external systems. Each skill defines a uniform interface for actions such as database queries, file reads, API calls, document collaboration, PDF handling, Excel automation, and Slack messaging.

The repository serves as a reference implementation for developers building Model‑Tool Communication Protocol (MCP)‑based agents, illustrating how to expose external tools as model‑understandable commands.

Open source address : https://github.com/anthropics/skills

5. Chatterbox – Low‑Latency Real‑Time Voice Dialogue Framework

Chatterbox, released by Resemble AI, eliminates the sluggish feel of voice interactions by streaming audio synthesis while the language model is still generating text. It integrates ASR, LLM, and TTS modules into a single pipeline.

The Turbo model uses a 350M‑parameter lightweight architecture, collapsing the typical ten‑step voice generation process into a single step, reducing memory usage and compute while preserving high‑fidelity audio. The framework also supports auxiliary tags like [cough] and [laugh] for more natural speech.

Chatterbox is ideal for virtual humans, intelligent customer‑service bots, or any AI assistant that requires real‑time, high‑quality voice output without dealing with low‑level streaming synchronization.

Open source address : https://github.com/resemble-ai/chatterbox

6. iptv‑org/iptv – Massive Public IPTV Channel Collection

The iptv‑org/iptv repository aggregates thousands of free, legally public IPTV channels from around the world, boasting over 108k GitHub stars and the most comprehensive maintenance among similar projects.

Channels are meticulously categorized by genre, language, and country, and provided as ready‑to‑use .m3u playlists compatible with VLC, PotPlayer, and other IPTV players. Automated scripts continuously verify links, removing dead or illegal streams to ensure reliability.

This collection is a valuable resource for developers building IPTV applications or users seeking diverse, free broadcast content.

Open source address : https://github.com/iptv-org/iptv