Daily Digest – 2026-07-20

vuejs/core

Progressive JS framework that scales from widgets to full SPAs without the bloat. (TypeScript, 54k★)

★ 54,185

vlang/v

Compiled language that bootstraps itself in <1s, auto-translates C, no runtime dependencies. (V, 37.7k★ ⚡15★/day)

★ 37,737

DioxusLabs/dioxus

Write web, desktop, mobile from one Rust codebase using React-like syntax. (Rust, 37.2k★)

★ 37,247

immich-app/immich

Self-hosted photo vault with mobile sync, no cloud vendor lock-in or monthly fees. (TypeScript, 108.3k★ ⚡113★/day)

★ 110,125

hyprwm/Hyprland

Wayland tiling compositor for Linux that looks polished without eating your RAM. (C++, 36.8k★)

★ 37,784

schollz/croc

Send files between machines securely without thinking about IPs, ports, or infrastructure. (Go, 36.2k★ ⚡104★/day)

★ 39,505

dashed/whisperx-subtitles-replicate

Generate accurate subtitles from audio faster than Whisper, hosted ready. (152K runs)

myaiteam2/websitescraper2

Extract emails, phone numbers, links from websites at scale via API. (130K runs)

google/nano-banana-2-lite

Ultra-fast image generation (text-to-image) optimized for speed over quality. (102K runs)

sync/lipsync-2-pro

Professional mouth-sync for talking head videos without expensive video editors. (53K runs)

veed/fabric-1.0

Turn static images into talking videos, useful for product demos and ads. (48K runs)

starling-lm

7B model trained on AI feedback, competitive with larger models on reasoning. (942.2K pulls)

phind-codellama

34B code generation model, faster than GPT-4 for most coding tasks locally. (939.2K pulls)

solar

10.7B model optimized for single-turn chats, compact and practical. (937.2K pulls)

yarn-llama2

Llama 2 extended to 128k context window, handles long documents locally. (931.5K pulls)

paraphrase-multilingual

Semantic embeddings across 50+ languages, for clustering and cross-lingual search. (929.1K pulls)

@appium/logger

Structured logging for Appium test automation across iOS, Android, web. (1.7M/wk)

@genkit-ai/core

Google's framework for building AI apps with type-safe flows and tracing. (263K/wk)

@genkit-ai/ai

Genkit APIs for working with LLMs, embeddings, and generative models. (257K/wk)

genkit

Full Genkit AI framework with local dev, Firebase integration, observability. (249K/wk)

@genkit-ai/firebase

Deploy Genkit AI flows to Firebase Cloud Functions with Firestore tracing. (230K/wk)

qdrant-client

Python client for Qdrant vector DB, handles RAG and semantic search. (4.9M recent downloads)

ollama

Run local LLMs via Python, no need to manage separate Ollama server. (4.5M recent downloads)

docling

Parse PDFs, Word docs, HTML into structured markdown for AI pipelines. (4.5M recent downloads)

instructor

Get structured JSON outputs from LLMs reliably without prompt wrestling. (4.4M recent downloads)

gradio

Build shareable web UIs for ML models in pure Python, no JS required. (3.3M recent downloads)

openai/whisper

Speech-to-text that handles accents and noise without vendor lock-in, runs locally. (gradio, 2796 likes)

★ 106,030

coqui/xtts

Text-to-speech that clones voices from short clips, multilingual, no API fees. (gradio, 2765 likes)

tonyassi/voice-clone

Clone voices from audio samples and generate speech, runs in Gradio without setup. (gradio, 2661 likes)

multimodalart/qwen-image-multiple-angles-3d-camera

Generate 3D object images from multiple angles using Qwen vision model. (gradio, 2622 likes)

fffiloni/diffusers-image-outpaint

Extend image boundaries intelligently, fills in context beyond original crop. (gradio, 2536 likes)

Qwen/Qwen3-4B

4B parameter model dense enough for local inference, runs on laptops. (text-generation, 5.0M downloads)

Qwen/Qwen2.5-0.5B-Instruct

Half-gig model that fits on edge devices, practical for offline-first apps. (text-generation, 4.8M downloads)

google/vit-base-patch16-224

Vision transformer for image classification, pre-trained and efficient. (image-classification, 4.5M downloads)

dphn/dolphin-2.9.1-yi-1.5-34b

34B parameter model tuned for reasoning, runs on high-end consumer hardware. (text-generation, 4.5M downloads)

mixedbread-ai/mxbai-embed-large-v1

Embedding model for semantic search and RAG, works offline with ONNX export. (feature-extraction, 4.4M downloads)

genrobot2025/10Kh-RealOmin-OpenData

1TB+ robotics dataset with real robot video for reinforcement learning training. (680K downloads)

XDOF/ABC-130k

Large-scale robotic manipulation dataset with 130k trajectories for training. (679K downloads)

mvp-lab/LLaVA-OneVision-1.5-Mid-Training-85M

Multimodal training dataset for vision-language models at scale. (652K downloads)

Fuzzy AI

We warm your prospects before reaching out (430 upvotes) [producthunt]

Nautis

The AI-native Operating System for founders. (366 upvotes) [producthunt]

Replay QA

Replay QA tells you what is broken before your users do (291 upvotes) [producthunt]

Skippr AI

The live AI employee inside your product, serving every user (257 upvotes) [producthunt]

Loova Ads Studio

Create AI ads that convert, not just look good (239 upvotes) [producthunt]

IKEA Complexity Index

Show HN: IKEA Complexity Index (178 upvotes) [hackernews_showhn]

LayerProof Mylar

Create launch-ready motion videos for your products (137 upvotes) [producthunt]

Creed

Your personal context file for every agent. (131 upvotes) [producthunt]