Daily Digest – 2026-08-05

OpenBB-finance/OpenBB

Free financial data APIs for quants and AI agents without the Bloomberg terminal bill. (Python, 71.4k★)

★ 71,430

k3s-io/k3s

Kubernetes in under 40MB—actually practical for edge, CI/CD, and resource-constrained deployments. (Go, 33.7k★)

★ 33,661

tokio-rs/tokio

Rust async runtime that actually lets you write concurrent code without losing your mind. (Rust, 32.8k★ ⚡35★/day)

★ 32,821

sundowndev/phoneinfoga

Scrape publicly available phone number data from online sources in bulk. (Go, 17.3k★ ⚡29★/day)

★ 17,340

apache/doris

Real-time OLAP database optimized for AI agent queries and hybrid search workloads. (Java, 15.7k★)

★ 15,723

ultralytics/yolov8s-worldv2

Open-vocabulary object detection on Replicate—detect any class you name without retraining. (17K runs)

black-forest-labs/flux-3

Text-to-video with synchronized audio via Replicate—early access to multimodal generation. (793 runs)

deepcoder

14B open-source coder model at O3-mini level—runs locally via Ollama. (937.9K pulls)

reader-lm

HTML-to-Markdown converter models—14B and 1.5B variants for content extraction. (907.5K pulls)

shieldgemma

Safety classifier for prompts and responses—evaluate LLM outputs against policies. (907K pulls)

llama-pro

Llama 2 extended with code and math knowledge—hybrid general and domain expertise. (889.9K pulls)

yarn-mistral

Mistral extended to 64K or 128K context—long-document reasoning without API costs. (882.7K pulls)

@opentelemetry/instrumentation-openai

Auto-instrument OpenAI API calls for traces and latency metrics in Node apps. (5.9M/wk)

@tanstack/ai-client

Headless chat client for streaming, realtime sessions, and media generation—framework agnostic. (143K/wk)

genkit-cli

CLI for building and testing flows in Google Genkit without leaving terminal. (94K/wk)

@genkit-ai/telemetry-server

Local telemetry server for Genkit flows—debug traces and latency locally. (90K/wk)

repomix

Pack entire repos into single file for pasting into LLMs—saves token math. (87K/wk)

crewai

Multi-agent framework where agents collaborate on tasks—think team of workers, not one bot. (5.3M recent downloads)

qdrant-client

Official Python client for Qdrant vector DB—semantic search and RAG backbone. (4.5M recent downloads)

outlines

Generate structured outputs (JSON, regex) from LLMs without prompt engineering hacks. (705K recent downloads)

together

Official Together.ai Python SDK—run open-source models cheaper than proprietary APIs. (655K recent downloads)

selfit-camera/Omni-Image-Editor

Browser-based image editor with region selection and editing via Gradio—no backend needed. (gradio, 2316 likes)

prithivMLmods/Qwen-Image-Edit-2511-LoRAs-Fast

Fast image editing with Qwen LoRAs—runs inference locally via Gradio. (gradio, 2261 likes)

jbilcke-hf/FacePoke_CLONE-THIS-REPO-TO-USE-IT

Face manipulation tool via Docker—clone, run, edit faces locally without external APIs. (docker, 2214 likes)

tencent/Hunyuan3D-2.1

Text-to-3D model accessible via Gradio—generates 3D assets from prompts. (gradio, 2175 likes)

finegrain/finegrain-image-enhancer

Upscaler and image enhancer leveraging Stable Diffusion—live demo in browser. (gradio, 2161 likes)

ornith-ai/Ornith-1.0-9B-GGUF

9B conversational model in GGUF format—quantized and runnable on commodity hardware. (text-generation, 4.9M downloads)

jonatasgrosman/wav2vec2-large-xlsr-53-russian

Pre-trained Russian speech recognition model—fine-tune or use as-is for transcription. (automatic-speech-recognition, 4.0M downloads)

Qwen/Qwen3-Embedding-8B

8B text embedding model optimized for semantic search and RAG retrieval. (feature-extraction, 3.3M downloads)

jinaai/jina-embeddings-v3

Multilingual embedding model with strong MTEB scores—supports ONNX and PyTorch. (feature-extraction, 3.2M downloads)

timm/resnet50.a1_in1k

Pre-trained ResNet50 from timm library—production-ready image classification backbone. (image-classification, 2.8M downloads)

applied-ai-018/pretraining_v1-omega_books

100M+ book records in Parquet format for pre-training or fine-tuning at scale. (726K downloads)

Maximilians/ps2_hf1

Dataset for training models—check repo docs for task details and sample composition. (603K downloads)

allenai/ai2_arc

Multiple-choice QA dataset from science exams—benchmark for reasoning model evaluation. (451K downloads)

Simple algorithm and color space to generate diverse skin tones

Show HN: Simple algorithm and color space to generate diverse skin tones (586 upvotes) [hackernews_showhn]

AdAnt AI

Claude for viral, high-converting social ads (311 upvotes) [producthunt]

Wispr Flow Notetaker

Meeting notes that get the details right. (296 upvotes) [producthunt]

ngrok AI Gateway

One private gateway for every AI model (240 upvotes) [producthunt]

Cloudflare Wallets

the programmable wallet for the agentic Internet (183 upvotes) [producthunt]

Maple-Preview – Ternary 20B MoE running at 120 tok/s on a iPhone

Show HN: Maple-Preview – Ternary 20B MoE running at 120 tok/s on a iPhone (158 upvotes) [hackernews_showhn]

Kiro Crew

Open source agentic development workspace (151 upvotes) [producthunt]

Fine-tune an 8B model on a 4 GB laptop GPU

Show HN: Fine-tune an 8B model on a 4 GB laptop GPU (132 upvotes) [hackernews_showhn]