PyTorchKR news
The latest AI and developer news from PyTorchKR, with short summaries.
Visit website →Pickimg2threejs Released, Converting Images into Three.js Code
It is an AI agent tool that analyzes a single image and reconstructs it into 3D model code based on Three.js.
Zilliz Unveils 'memsearch', a Shared Memory Tool for AI Agents
It's a tool that provides a markdown-based persistent memory layer so multiple AI coding agents can share context.
PickOpen-Source Leaderboard for Korean Embedding Models Released
An open-source embedding leaderboard has launched, enabling objective comparison of Korean RAG and retrieval performance.
Helion, a High-Performance TPU Kernel DSL for PyTorch, Released
Helion, a high-level DSL that lets developers write high-performance TPU kernels in PyTorch style, has been released.
PickGigatoken released, processing gigabytes per second
gigatoken, a Rust-based tokenizer offering up to 1000x faster speed than HuggingFace tokenizers, has been released.
AI Agent-Based Security Research Platform open-kritt Released
It is an open-source security platform that breaks down complex security review tasks into small units, which multiple AI agents then perform in parallel.
Crawlee, a Data Collection Library for AI/LLM, Released
Introducing Crawlee, a high-performance Node.js web scraping library for building LLM and RAG pipelines.
PostgreSQL-based AI Search Engine, pgContext, Released
pgContext, an extension that supports both vector search and hybrid search within PostgreSQL, has been released.
Local-based Web Search MCP Server 'wigolo' Released for AI Agents
It is an MCP server tool for AI agents that performs web search and crawling locally without the burden of API key costs.
PyTorch Foundation Announces Updates on 6 Major Projects
The PyTorch Foundation revealed the latest achievements and roadmaps for 6 projects, including PyTorch and vLLM.
OpenAI Unveils Secure MCP Tunnel
OpenAI has unveiled Secure MCP Tunnel, which connects private MCP servers to ChatGPT and other services without exposing them externally.
Self-Evolving AI Agent Phantom Unveiled
Introducing Phantom, an AI agent that builds its own infrastructure, learns, and evolves through a dedicated VM.
OfficeCLI, an Office Tool for AI Agents, Released
OfficeCLI, a CLI tool that lets AI agents read, modify, and visually verify MS Office documents, has been launched.
LoopGain Cuts AI Agent Loop Costs by 92%
LoopGain, an open-source library based on control theory, automatically terminates repetitive AI agent loops to cut API costs by up to 92.8%.
Perplexity Releases WANDR Benchmark
Perplexity has released WANDR, an open benchmark that evaluates a research agent's ability to collect data at scale.
Pickkordoc, an Open-Source MCP Server for Converting Korean Documents to Markdown, Released
kordoc, an open-source tool that converts HWP, PDF, and Office documents to Markdown and integrates with AI agents as an MCP server, has been released.
Zyphra Releases Open EEG Foundation Model ZUNA1.1
Zyphra has released ZUNA1.1, an open foundation model under Apache 2.0 that restores, denoises, and upsamples EEG signals from arbitrary channel layouts.
Pickfeyn releases text-to-SQL model SQRL
feyn, the developer of Chonkie, has released the SQRL model series under Apache 2.0, which generates SQL after directly exploring the DB.
pi-computer-use, GUI Automation Agent Extension Released
pi-computer-use, an MIT open-source extension that lets AI agents directly operate desktop apps without APIs, has been released.
Reasoning Expands the Boundary of LLM Factual Recall Ability
A joint research team including Google Research revealed in a COLM 2026 paper that activating reasoning even for simple factual questions expands the boundary of LLM parametric knowledge recall.