What changed in Tech today
Articles, releases and code from Hacker News, Reddit, GitHub and the people building RPA, workflow automation and AI agents — plus what the community pushed to the top today.
Not enough to go on yet. Open a few links and this list will start leaning towards what you read. 0 of 5 so far.
-
01
microsoft/autogen — A programming framework for agentic AI
GitHubgithub.com 61657 pts +15/h 16 Oct
-
02
DayuanJiang/next-ai-draw-io — Draw and edit draw.io diagrams by chatting with an AI. Web app, desktop app, and an MCP server for Claude Code, Cursor, Codex and VS Code.
AIThe open-source Next AI Draw.io project lets users create and edit diagrams via conversational AI, offering an MCP server to integrate diagramming directly into coding agents.
GitHub -
03
windmill-labs/windmill — Code-first platform for workflows, apps and AI agents, on the fastest self-hostable orchestrator (benchmarked against Airflow, Prefect and Temporal). Business teams build with AI, engineers control what ships. Open-source alternative to Lovable, Retool and n8n.
GitHub -
04
n8n-io/self-hosted-ai-starter-kit — The Self-hosted AI Starter Kit is an open-source template that quickly sets up a local AI environment. Curated by n8n, it provides essential tools for creating secure, self-hosted AI workflows.
AIn8n released a Docker Compose starter kit combining n8n, Ollama, Qdrant, and PostgreSQL, letting developers deploy fully local, privacy-compliant AI automation workflows.
GitHub -
05
autotrust/JEV-27B-VL — image-text-to-text
AIThe JEV-27B-VL vision-language model enables fast, single-pass visual decision-making, allowing UI automation and robotic agents to select actions directly from screenshots in roughly 240 milliseconds.
HFhuggingface.co 6978 pts +28/h 5 Oct
-
06
facebook/sam3 — mask-generation
HF -
07
baidu/Unlimited-OCR — image-text-to-text
HF -
08
kajisho5/ffmpeg-skill — Local video editor for coding agents: 42 FFmpeg tools with a machine-readable contract and an MCP server, for Claude Code, Cursor and Codex. Cut, captions, reframe, loudness, platform exports. No cloud, no API keys.
AIA new open-source MCP server and skill package exposes 42 local FFmpeg tools to coding agents, enabling video editing workflows without external APIs or cloud services.
GitHub -
09
deepseek-ai/DeepSeek-V4.1-Flash-1017 — image-text-to-text
HF -
10
pyannote/speaker-diarization-community-1 — automatic-speech-recognition
HF -
11
cranebeaker/coop-anything — Co-op Anything: point your AI agent at any single-player game and play it with friends (online and/or splitscreen!). MCP server and playbook to let you reverse engineer, fix, test, ship and share what it learns.
AICo-op Anything provides an MCP server and playbook that enables AI coding agents to reverse engineer, develop, and test multiplayer mods for single-player video games.
GitHub -
12
hujjun/mcd-saver-strategist — 麦当劳省钱点餐助手（麦门精算师）｜麦当劳优惠券组合优化 Skill：拉取麦当劳菜单与全部可用优惠券，算出最省的点餐组合，巨无霸套餐 ¥47 → ¥21.9（省 53%）。McDonalds MCP（麦当劳 MCP）coupon optimizer · McDonalds China MCP Server · Agent Skill for WorkBuddy · 麦当劳程序员创意开发大赛参赛作品
AIAn open-source agent skill connects WorkBuddy to the McDonald's China MCP server, letting developers automate coupon optimization, price calculation, and order placement via programmatic APIs.
GitHub -
13
ollaya-dev/ollaya — Jev-compatible API server for open decision models. Ollama for decision models: pull and serve 24 model families (Laya, pplx-decider, Clef, Winnow, Kev...) locally or self-hosted behind /v1/systemone. TypeSafe SDK works by changing one env var. CLI, daemon, desktop app, Docker, MCP server, public bench.
AIOllaya released an open-source API server to locally serve decision models, allowing developers to run specialized decision-making logic locally via an MCP server or SDK.
GitHub -
14
HexRaysSA/ida-mcp — Official Hex-Rays IDA MCP Server.
AIHex-Rays released an official IDA MCP server, enabling AI agents to automate binary analysis workflows in IDA Pro via stdio or HTTP interfaces.
GitHub -
15
ASCIT31/Dark-Moon — Open-source autonomous AI penetration testing. 50 specialist agents across web, API, cloud, Active Directory, Kubernetes, CI/CD and AI/LLM (OWASP LLM Top 10). Runs on a local LLM behind a Privacy Gateway. Integrates with GitHub, GitLab, Jenkins, VS Code, JetBrains, n8n, Grafana & Splunk.
GitHub -
16
bytechefhq/bytechef — Open-source AI agents and workflow automation. Self-host or embed in your SaaS. Apache 2.0 alternative to n8n and Zapier.
AIByteChef provides an Apache 2.0 open-source platform to build, self-host, and embed AI agents and workflow automation directly into SaaS applications.
GitHub -
17
hanzohanzhe/Grandet — Find the cheapest API for any model and connect it from your AI agent: Grandet Skill and an anonymous MCP server, backed by listed prices from 1,800+ official, inference and relay providers on grandet.ai.
AIGrandet released an MCP server and skill that lets AI agents dynamically query model pricing across 1,800 providers to automatically route tasks to the cheapest API.
GitHub -
18
alexgaoth/awesome-reverse-engineered — Don't reverse engineer it twice. 16,000+ open-source tools built by reverse engineering closed apps, APIs, devices, protocols, file formats, games and AI tools. Search site + agent skill + MCP server, refreshed nightly.
AIAwesome-reverse-engineered cataloged over 16,000 open-source tools with an MCP server and agent skill, helping builders integrate closed APIs, devices, and protocols into automated workflows.
GitHub -
19
LiCinCan/blender-ai-mcp-addon — AI assistant and MCP server for Blender: visual chat, scene tools, multi-agent workflows. Works with OpenAI, ChatGPT, Claude, Gemini, Ollama, OpenCode. Blender AI addon and plugin, copilot for 3D, bpy Python scripting, MCP integration.
AIAn open-source Blender add-on introduces a built-in Model Context Protocol server, enabling external AI agents and LLMs to inspect scenes, run Python scripts, and automate 3D modeling tasks.
GitHub -
20
scenario-labs/blender-plugin — Scenario for Blender: AI images, video, 3D and PBR materials in the viewport, plus a local MCP server for agents. Blender 5.0+ extension, GPL-3.0-or-later.
AIA new Blender extension embeds generative AI tools directly into the viewport and includes a local Model Context Protocol server, enabling AI agents to automate 3D workflows.
GitHub -
21
windmill 1.824.0
AIWindmill 1.824.0 introduces an AI decision flow step, adds a Notion OAuth provider, and expands Google and Slack integration scopes.
Release -
22
kestra 2.0.5
AIKestra 2.0.5 introduces a centralized HTTP retry mechanism, adds configurable worker job buffers, and fixes several execution retry, loop concurrency, and breakpoint handling issues.
Release -
23
n8n n8n@3.0.0
AIn8n 3.0.0 fixes agent node credential access, strengthens streaming and execution security, and enhances AI assistant and task runner stability across automated workflows.
Release -
24
kestra 2.0.8
Release -
25
windmill 1.832.0
AIWindmill 1.832.0 tracks Python versions per instance and defaults new ones to Python 3.13, while improving variable storage capacity and S3 proxy security.
Release -
26
kestra 2.0.6
AIKestra 2.0.6 converts the Schedule trigger away from polling and resolves bugs with subflow retries, task duplication, and looped executions.
Release -
27
trigger.dev .docker.4.7.2: chore: release v4.7.2 (#4996)
AITrigger.dev v4.7.2 fixes ordered tool results for chat agents to preserve model reasoning, prevents worker timeouts during long response recovery, and sets default Node.js LTS runtimes.
Release -
28
prefect 3.9.0 - Pendulum's last swing
AIPrefect 3.9.0 drops Python 3.10 support, deprecates Pendulum datetimes, removes legacy runner paths, and improves automation resilience during websocket outages.
Release -
29
trigger.dev .docker.4.8.0: chore: release v4.8.0 (#5044)
AITrigger.dev v4.8.0 adds hosted webhooks and agent chat routing, letting developers automatically verify inbound events, pause runs until webhooks arrive, and connect AI agents to external chat channels.
Release -
30
langgraph langgraph-sdk==0.4.9
AILangGraph SDK 0.4.9 adds a Priority feature for thread settings, allowing developers to manage the execution precedence of concurrent agent workflows.
Release -
31
langgraph langgraph==1.2.13
AILangGraph 1.2.13 fixes checkpoint replay and branching issues, improving state management and interrupt tracking reliability for complex multi-agent workflows.
Release -
32
airflow providers-duckdb/0.2.0
Release -
33
airflow providers/2026-10-02
Release -
34
mcp servers 1.0.0
AIModel Context Protocol reference servers reached version 1.0.0, providing stable packages for filesystem, memory, and task execution integrations in AI agent workflows.
Release -
35
temporal 1.33.0-164.2
Release -
36
google/embeddinggemma-2 — feature-extraction
HF -
37
autotrust/GEV-26B-Decide — text-classification
AIAutoTrust released GEV-26B-Decide, an open-weights decision model that enables sub-100ms computer use and tool selection by combining fast single-pass inference with adaptive reasoning.
HF -
38
mistralai/Mistral-Large-4-1T-A52B — image-text-to-text
HF -
39
RecursiveSI/flint — text-generation
AIRecursive released Flint, an open-weights 40B mixture-of-experts model optimized for coding agents, multi-turn tool use, and codebase automation with 2.5B active parameters per token.
HF -
40
openbmb/MiniCPM-V-4.7-35B-A3B — image-text-to-text
HF -
41
Qwen/Qwen-Image-2.1-Turbo — text-to-image
AIQwen released Qwen-Image-2.1-Turbo, an accelerated 7B text-to-image model using eight denoising steps, enabling faster, lower-latency image generation inside Diffusers-based automation pipelines.
HF -
42
XHToken/Spark-X2.6-4B — text-generation
HF -
43
Aleph-Alpha/Kolibri-1 — text-generation
AIAleph Alpha released Kolibri-1, an open-weight 78B bilingual MoE model with native tool calling and long-context support for German and English agentic automation workflows.
HF -
44
Nanbeige/Nanbeige4.3-33B — text-generation
HF -
45
Cloudflare/clef-flash — image-text-to-text
AICloudflare released Clef-Flash, a 9B multimodal model that evaluates structured schemas directly into decision probabilities, eliminating text generation and output parsing for agent workflows.
HF -
46
-
47
Venastine-Research/Xing4.0-29B-A4B-GGUF — text-generation
HF -
48
meta-models/Muse-Spark-Open-816B-A42B — image-text-to-text
HF -
49
anujasid/blender-ai-mcp-addon — AI assistant and MCP server for Blender: visual chat, scene tools, multi-agent workflows. Works with OpenAI, ChatGPT, Claude, Gemini, Ollama, OpenCode. Blender AI addon and plugin, copilot for 3D, bpy Python scripting, MCP integration.
AIA new Blender add-on embeds an MCP server, enabling AI agents and external clients to automate 3D scene creation, material editing, and asset workflows.
GitHub -
50
LiquidAI/LFM3-1.7B — text-generation
HF -
51
graygnatconsole/mcp-audit-tool — 🛡️ Security audit CLI for Model Context Protocol (MCP) servers — scan AI agent configs for tool poisoning, rug pulls, hardcoded secrets, command injection & supply-chain risks. Pure Python, SARIF + CI ready.
AIA new open-source CLI audits Model Context Protocol client configurations to detect security vulnerabilities, command injection risks, hardcoded secrets, and tool poisoning before running AI agents.
GitHub -
52
ComputerUseAgent/AutoCua — Automatic computer use agent for end-to-end automation workflows on macOS, Windows, Linux, iOS and the browser.
GitHub -
53
SC117/Qwen3.8-Flash-Next-GSQ-RCO-abliterated-GGUF — image-text-to-text
HF -
54
LiquidAI/LFM2.5-24B-A2B — text-generation
HF -
55
autotrust/GLM5.3-Flash-E224-DGX-Spark — image-text-to-text
AIA pruned NVFP4 build of GLM-5.3-Flash enables developers to run high-accuracy, vision-capable reasoning models and agent workflows on dual desktop NVIDIA DGX Spark systems.
HF -
56
nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B — automatic-speech-recognition
AINVIDIA released Nemotron-3.5-Transcribe-30B-A4B, an open-weight speech model supporting 40 languages, enabling developers to build fast, local audio automation pipelines with speaker attribution and context understanding.
HF -
57
Comfy-Org/MOSS-Image
AIComfy-Org released repackaged MOSS-Image model weights, allowing builders to integrate the OpenMOSS image generation and fine-tuned models directly into ComfyUI automation workflows.
HF -
58
MiniMaxAI/MiniMax-M3.1-Flash — image-text-to-text
HF -
59
ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF — text-generation
HF -
60
baidu/Unlimited-OCR-2 — image-text-to-text
HF -
61
datalab-to/indra-ocr — image-to-text
HF -
62
tencent/Hy-SimulMT-30B-A3B-FP8 — translation
HF -
63
LiquidAI/d1-3B — image-text-to-text
HF -
64
KnifeLemon/Droidline — Android automation from code. Test your apps on real phones, script the tasks you repeat, run a shelf of phones, or let an AI agent use one, from Python, Node.js, the CLI, HTTP or Appium clients. No ADB, no USB cable, no root.
AIDroidline released an open-source Android automation tool that lets developers control real devices using Python, Node.js, or AI agents without requiring ADB, USB cables, or root access.
GitHub -
65
mistralai/Timestral-209M — time-series-forecasting
HF -
66
ccore/Qwen3.5-4B-3D-Coder
AICCORE released an early preview of Qwen3.5-4B-3D-Coder, enabling developers to automate the generation of interactive Three.js 3D web scenes using a lightweight local model.
HF -
67
openbmb/MiniCPM-V-4.7-1B — image-text-to-text
HF -
68
orcarouter/DeepSeek-V4.1-Flash-1017-abliterated-GGUF — text-generation
AIOrcaRouter released GGUF quantizations of an abliterated DeepSeek-V4.1-Flash model, letting developers run an uncensored 763-billion-parameter architecture locally for automated security testing and red-teaming workflows.
HF -
69
openbmb/MiniCPM-OCR-1.0 — image-text-to-text
HF -
70
InternScience/Sage-A1.5 — text-generation
HF -
71
unsloth/flint-GGUF — text-generation
HF -
72
nvidia/Qwen3.8-Flash-Next-FP4-Local — image-text-to-text
HF -
73
nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB — text-generation
AINVIDIA released Nemotron-3.5-Bolt-E4B-4GB, an open 4-billion parameter language model optimized for multi-turn tool use, fast local inference, and low-memory execution in automated edge agents.
HF -
74
tencent/Hy-Memory-30B-A3B — text-generation
HF -
75
unsloth/embeddinggemma-2-GGUF — feature-extraction
AIUnsloth released GGUF quantizations of Google's EmbeddingGemma 2, enabling developers to run lightweight, multimodal text, image, audio, and video embedding pipelines locally for RAG and search.
HF -
76
Pika-Labs/Pika-H3-Flex — image-to-video
AIPika released the open-weight Pika-H3-Flex model, letting developers integrate fast six-step video and synchronized audio generation with multi-reference control directly into automated media workflows.
HF -
77
orcarouter/OrcaCyber-Nano-1.0 — image-text-to-text
AIOrca released OrcaCyber Nano 1.0, a 27-billion-parameter vision-language model in GGUF format that lets developers run local, air-gapped cybersecurity triage and code review agents.
HF -
78
LiquidAI/LFM3-Omni-2B — any-to-any
AILiquid AI released LFM3-Omni-2B, a 2.2-billion-parameter multimodal model that lets developers run end-to-end voice, vision, and text automation workflows locally on edge devices.
HF -
79
reflection/Beam-1.0-501B-A23B — text-generation
HF -
80
-
81
LiquidAI/LFM2.5-24B-A2B-GGUF — text-generation
HF -
82
Ablation-Tool/ablation — Ablation is a reverse engineering framework built for autonomous operation with AI coding agents
AIAblation has launched an open-source reverse engineering framework designed for autonomous operation by AI coding agents, allowing developers to automate software analysis tasks.
GitHub -
83
KRAFTON/Raon-1-mini-preview — text-generation
HF -
84
Comfy-Org/Fibo-2.0
HF -
85
nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4 — image-text-to-text
AINVIDIA released Nemotron-3.5-Super-VL-120B-A12B-NVFP4, a quantized vision-language model supporting 1M tokens, enabling builders to run multimodal reasoning and tool-calling agents on fewer GPUs.
HF -
86
lodestones/MOSS-Image-SFT-8-steps-dmd
AIA new distillation of MOSS-Image-SFT cuts generation to eight steps without classifier-free guidance, enabling faster image synthesis in automated media workflows.
HF -
87
meta-models/Muse-Spark-Open-816B-A42B-NVFP4 — image-text-to-text
HF -
88
ornith-ai/Ornith-1.5-122B — text-generation
HF -
89
LiquidAI/LFM3-VL-2B — image-text-to-text
AILiquid AI released LFM3-VL-2B, a compact vision-language model enabling edge devices to run step-by-step reasoning on images, documents, and video inputs for automation workflows.
HF -
90
Diffuspace/Clover-35B-A3B-Preview — text-generation
AIDiffuspace released Clover-35B-A3B-Preview, an open-source diffusion language model that combines block-level parallel decoding with a 128K context window for fast, long-context text generation in automation pipelines.
HF -
91
briaai/Fibo-2.0-dev — text-to-image
HF -
92
nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 — text-generation
HF -
93
openbmb/MiniCPM-V-4.7-35B-A3B-gguf — image-text-to-text
HF -
94
FastVideo/FastH3-Omni-Comfy
HF -
95
-
96
autotrust/GEV-26B-Decide-NVFP4 — text-classification
AIAutotrust released an NVFP4-quantized version of GEV-26B-Decide, cutting GPU memory usage by 66% so developers can run fast decision-making models for UI automation on 24GB GPUs.
HF -
97
Edge0/Audio8-TTS-Realistic-0.6b — text-to-speech
HF -
98
LiquidAI/LFM3-Audio-2B — any-to-any
HF -
99
NaiveAI/Naive-N0.5-Flash — text-generation
AINaiveAI released Naive-N0.5-Flash, an open-weight 309B MoE model offering a 1M-token context window and fast inference for long-context coding and agent workflows.
HF -
100