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@descix/cli
DeSciX CLI and MCP Server - Unified command-line interface for DeSciX platform
- Tool@gnldev/rag
Vector store + RAG tool for AI SDK agents. Used as a tool, it becomes replayable/exactly-once RAG with @gnldev/durable.
- Tool@microsoft/agents-a365-observability
OpenTelemetry tracing and monitoring SDK for AI agents built with TypeScript/Node.js
- Toolclaude-code-cost
Cost dashboard for Claude Code usage
- Tool@microsoft/agents-a365-runtime
Agent 365 Runtime SDK for AI agents built with TypeScript/Node.js
- Tool@descix/sdk
DeSciX SDK - Shared runtime libraries for Node.js and Browser
- Tool@microsoft/agents-a365-tooling
Agent 365 Tooling SDK for AI agents built with TypeScript/Node.js
- Tool@runtypelabs/sdk
TypeScript SDK for the Runtype API with fluent methods. Use it to quickly realize AI products, agents, and workflows.
- Tool@microsoft/agents-a365-tooling-extensions-langchain
Agent 365 Tooling SDK for LangChain AI agents built with TypeScript/Node.js
- Tool@microsoft/agents-a365-observability-extensions-langchain
Microsoft Agent 365 SDK LangChain instrumentation extensions for telemetry and observability
- Tool@bastani/pi-ai
Bastani-branded unified LLM API with automatic model discovery and provider configuration
- Tool@takara-ai/miru-code
Miru (見る) — hybrid code search for agents. Bun, TypeScript, Takara embeddings.
- Tooljg-utilities
Validations, transformers, generators, and collection, async, and environment utilities for browsers and Node.js.
- Tool@archmax-ai/harness
Agents can't reliably follow a process. archmax harness makes them: a governed layer over LangChain Deep Agents that enforces a declarative workflow state machine on ever
- Toolburgee
A CLI framework built on one declaration: help, --json, --schema, MCP, completions and types are projected from it, with exit codes and fix: lines an agent can act on. No
- Toolseniority
Resolves flags, environment variables, config files and defaults in one fixed order, with provenance: every value can say where it came from. Zero dependencies. Drop-in p
- Tool@ashx-j/lunr
lunR — terminal coding agent CLI with read, bash, edit, write tools and session management
- Tool@ashx-j/lunr-ai
Unified LLM API with automatic model discovery and provider configuration
- Toolpolotno
Design editor SDK: embed a Canva-like editor, generate designs from templates and data, export PNG, PDF, and video in the browser or on a server.
- Toole2b
E2B SDK that give agents cloud environments
- Tool@circuitwall/jarela
Jarela — local chat interface for LangGraph agents (multi-provider, single-process, SQLite-backed).
- Tool@e2b/cli
CLI for managing e2b sandbox templates
- Tool
- Tool@salesforce/b2c-dx-mcp
MCP server for B2C Commerce developer experience tools
- Toolforgedock
Autonomous development pipeline for coding agents. GitHub as institutional memory, with native Claude Code and OpenCode entrypoints.
- Tool@trillboards/edge-federated
On-device federated learning, gradient upload, and VAS attestation for Trillboards Edge AI SDK
- Tool@moss-dev/moss-web
Moss browser SDK — in-browser semantic search powered by WASM
- Tool@struktoai/mirage-core
A Unified Virtual Filesystem For AI Agents
- Tool@damphuquy/agent-init
Scaffolding CLI to bootstrap RIPER-5 Coding Agents & Operational Workspace
- Toolmcp-remote
Remote proxy for Model Context Protocol, allowing local-only clients to connect to remote servers using oAuth
- Tool@taiga-ai/mcp-server
Taiga for AI coding assistants: read products, policies, initiatives and runs, and operate the line, over MCP
- Tool@vercel/agent-eval
Framework for testing AI coding agents in isolated sandboxes
- Tool@dfm-fi/agent
DFM v2 MCP server — drive the full fund launch → deposit → redeem (+ manage) lifecycle from an AI agent against the live API.
- Toolaer-1
AER-1: verifiable receipts for AI agents (JavaScript/TypeScript). One checkable AI agent receipt per tool call, verifiable offline with no server, in the open AER-1 forma
- Tool@oneshot-agent/mcp-server
MCP server for OneShot - 173 tools for AI agents
- Toolprompt-protection
Provenance-tracked tool-call guard for JS agents (CaMeL-style), fail-closed, zero dependencies. Taints tool results, trusts what the user named, and blocks untrusted data
- Tool@primeng/mcp
Model Context Protocol (MCP) server for PrimeNg component library
- Tool@struktoai/mirage-node
A Unified Virtual Filesystem For AI Agents
- Tool@intangle/mcp-server
Model Context Protocol server for Intangle - AI context that persists across conversations
- Tool@s8fy/xdoc
Local PowerPoint CLI and MCP server for editable PPTX, native 3D charts, JSON patches, template data binding, PDF/PNG export and Markdown extraction for RAG.
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.