New
Everything new in AI as it appears: launches announced by the labs, brand-new models, datasets, papers and open-source tools. Checked every minute or two. Launches feed · Trending · News
Live now
- Connecting…
Tuesday 20 October
- Tool@venturekit-pro/ai
AI utilities for VentureKit - LLM gateway, agent loop, durable harness, embeddings, RAG, chat completion, and image generation
- Tool@vibeads/mcp
Official VibeAds MCP server: talk to your Google Ads account from Claude, Cursor or any MCP-compatible client with an API key (ChatGPT, Claude and Claude Code can also si
- Tool@pyxmate/memory
SDK for pyx-memory — Memory as a Service for AI agents
- Tool@uzen/paddleocr-js
Browser-based OCR SDK powered by PaddleOCR, ONNX Runtime Web and OpenCV.js
- Tool@opencraw/mcp
An MCP server exposing OpenCraw's crawler as agent tools: probe a page, validate recipes, run a crawl, list recipes.
- Tooldarce-cli
The coding agent you can undo: 250+ models in your terminal, parallel swarms, model racing, and /undo for everything, even shell commands.
- Tool@procedures/llm
A language model as procedures: a completion with tools, a stream of the answer and a structured answer, through the Anthropic Messages API, with the key from the environ
- Tool@procedures/client
Procedures as data for TypeScript: define typed procedures once and call them in process, over HTTP, WebSocket, the CLI or MCP. Streaming, middleware, retries, caching an
- Toolneuro-vault-mcp
MCP server for semantic and lexical search over Obsidian vaults, with a self-built embedding index.
- Tooln8n-nodes-digital-organizing
n8n community nodes for the tools Digital Organizing works with: Payrexx, RaiseNow, Cura Fundraising, Gravity Forms, Metricool, Webling and our own Do Counter, cevAPI, Ad
- Toolprivateer-agent
Privacy-first terminal coding agent — bring your own model across 20 providers (Anthropic, OpenAI, OpenRouter, Google, local Ollama…). Safe-by-default permissions, MCP, s
- Toolpostiz
Postiz CLI - Command line interface for the Postiz social media scheduling API
- Tool@docling/docling-core
TypeScript definitions and functions for using Docling output.
- Toolpondok-ai
Run your own private AI stack locally with one command.
- Tool@automattic/mcp-wordpress-remote
MCP WordPress Remote proxy server
- Tool@workerdeck/queue
Job queue over the WorkerDeck session runner: remote services schedule one-shot runs; the queue executes them as sessions with bounded concurrency and token budgets, deli
- Tool@workerdeck/web
The WorkerDeck dashboard as prebuilt static files: session list, create/resume flow, live panel, jobs, profiles, settings. Ships dist/ plus the path to it - serve it from
- Toolworkerdeck
Turnkey WorkerDeck instance: `npx workerdeck` runs the session gateway and the full dashboard on one port. The @workerdeck/* packages are the libraries you embed; this is
- Tool@workerdeck/client
Typed WorkerDeck protocol client for browsers and Node: REST session management plus a WebSocket attach with auto-reconnect and replay-from-last-seq. Uses the platform's
- Tool@workerdeck/server
The WorkerDeck gateway: HTTP + WebSocket session server over @workerdeck/core. Session registry (create/list/attach/interrupt/kill), pluggable auth hook, replay-from-seq
- Tool@workerdeck/react
Headless React layer for WorkerDeck: session hook + pure transcript and sessions-list reducers. No styling opinion - @workerdeck/ui is the styled layer on top.
- Tool@workerdeck/ui
Styled agent-control component library for WorkerDeck: session panel, transcript with streaming, tool-call cards, permission prompts, session list, composer. Tailwind v4
- Tool@workerdeck/core
The WorkerDeck session runner: drives Claude Code, OpenAI Codex and any AI SDK provider behind one Runner interface with a push-based input queue, promotes canUseTool cal
- Tool@workerdeck/protocol
The WorkerDeck wire protocol: typed session events, commands, and REST shapes shared by server and clients. Dependency-free, browser-safe. This protocol is the product bo
- Tool@dallay/agentsync
A fast CLI tool to sync AI agent configurations and MCP servers across Claude, Copilot, Cursor, and more using symbolic links.
- Tool@tempoxyz/mercator
Connect agent clients to the Mercator remote MCP server
- Tool@mcpjam/sdk
MCP server unit testing, end to end (e2e) testing, and server evals
- Tool@dingtalk-real-ai/dingtalk-connector
Official OpenClaw DingTalk channel plugin | 钉钉官方 OpenClaw 插件
- Tool@mcpjam/cli
Test, debug, and validate MCP servers — health checks, OAuth conformance, tool-surface diffing, and structured triage from the terminal or CI
- Toolpi-cache-optimizer
Improve Pi prompt/KV cache hit rates with stable prompts, OpenAI-compatible cache keys, proxy compat warnings, and footer cache stats.
- Toolvantadb
VantaDB TypeScript SDK — browser AI agent memory: WASM-powered embedded vector & graph memory for JS runtimes (Node, Bun, Deno, browser)
- Tool@node9/proxy
Access control for AI agents. Set what Claude Code, Codex, Gemini, Cursor and any MCP server are allowed to do, review risky actions before they run, and keep every actio
- Toolnode9-ai
Access control for AI agents. Set what Claude Code, Codex, Gemini, Cursor and any MCP server are allowed to do, review risky actions before they run, and keep every actio
- Tool@zhin.js/adapter-napcat
Zhin.js NapCat adapter for Plugin Runtime (WebSocket client, OneBot11 + NapCat extensions)
- Tool@zhin.js/adapter-telegram
Zhin.js Telegram Bot API adapter for Plugin Runtime (long-poll getUpdates)
- Tool@zhin.js/plugin-60s
60s API aggregation plugin for Zhin.js — news, weather, hot searches, utilities (Plugin Runtime)
- Tool@zhin.js/adapter-wechat-mp
Zhin.js WeChat Official Account adapter for Plugin Runtime (HTTP webhook)
- Tool@zhin.js/adapter-kook
Zhin.js KOOK adapter for Plugin Runtime (WebSocket gateway or Webhook)
- Tool@zhin.js/adapter-qq
Zhin.js QQ Official adapter for Plugin Runtime (WebSocket gateway)
- Tool@zhin.js/plugin-repeater
Repeater plugin for Zhin.js — auto-repeat when group members echo the same content (Plugin Runtime)
Headlines and links come from the labs' own blogs, Hacker News and the GDELT news index, and from the sources our index already follows. We show a short excerpt and always link to the original. A launch reported by a news site rather than the lab is marked “Reported” until the lab itself announces it.