Open Source Zapier with 380+ MCP Servers

Today’s top AI Highlights:

1. Open-source alternative to Zapier with 385 MCP servers

2. Agent-generated code from prompt to production

3. Anthropic has tripled its value in the last 6 months

4. A USB-C key that turns your phone into an AI assistant

5. Build full-stack apps with n8n-style visual workflows

& so much more!

Read time: 3 mins

AI Tutorial

We have created a complete Google Agent Development Kit crash course with 9 comprehensive tutorials!

This tutorial series takes you from zero to hero in building AI agents with Google's Agent Development Kit.

What's covered:

• Starter Agent - Your first ADK agent with basic workflow

• Model Agnostic - OpenAI and Anthropic integration patterns

• Structured Output - Type-safe responses with Pydantic schemas

• Tool Integration - Built-in tools, custom functions, LangChain, CrewAI, MCP

• Memory Systems - Session management with in-memory and SQLite storage

• Callbacks & Monitoring - Agent lifecycle, LLM interactions, tool execution tracking

• Plugins - Cross-cutting concerns and global callback management

• Multi-Agent Patterns - Sequential, loop, and parallel agent orchestration

Each tutorial includes explanations, working code examples, and step-by-step instructions.

Everything is 100% open-source.

We share hands-on tutorials like this every week, designed to help you stay ahead in the world of AI. If you're serious about leveling up your AI skills and staying ahead of the curve, subscribe now and be the first to access our latest tutorials.

Don’t forget to share this newsletter on your social channels and tag Unwind AI (X, LinkedIn, Threads) to support us!

Latest Developments

Open-source Alternative to Zapier with 385 MCP Servers 🕸️🧰

Zapier's pricing scales with usage until it becomes unaffordable, while most open-source automation tools are either feature-incomplete or require a PhD in infrastructure to deploy.

You're stuck paying premium prices for basic workflow automation or building everything from scratch.

Activepieces offers the best of both worlds: a fully-featured automation platform with 385 integrations that you can self-host or use in the cloud. The platform has a no-code canvas that lets non-technical users drag and drop automation flows, while developers can dive deep into custom pieces with full type safety and hot reloading. Every piece automatically becomes an MCP server compatible with Claude Desktop, Cursor, and Windsurf.

It has grown rapidly through community contributions, with over 60% of pieces built by users who wanted specific integrations and shared them with everyone else.

Key Highlights:

1. Largest MCP Ecosystem - The 385 pieces represent the biggest collection of open-source MCP servers available, with automatic compatibility across Claude Desktop, Cursor, and Windsurf. Each piece is battle-tested by thousands of users running millions of automated tasks monthly.

2. No-Code AI Agent Builder - Visual workflow designer with built-in AI Copilot assistance, plus native AI pieces for experimenting with different LLM providers. The platform includes an AI SDK for building custom agents that can request human approvals and handle complex multi-step processes.

3. TypeScript-First - All pieces are npm packages with full type safety and hot reloading support for local development. Developers can create custom integrations using modern TypeScript patterns and immediately share them with the community through the built-in package system.

4. Human-in-the-Loop - Workflows can request human approvals and handle back-and-forth conversations. The platform supports complex multi-step automations while keeping humans in control of important decisions.

Agent-Generated Code from Prompt to Production 💡📈

Warp, the agentic development environment, just shipped Warp Code with features that bridge the gap between agent-generated code and shipping it live.

The new release tackles the biggest pain point in AI coding: agents that write code that's "almost right, but not quite." According to Stack Overflow's 2025 survey, 66% of developers cite this as their top frustration. Warp Code introduces native code review for agent-generated changes, a lightweight file editor for quick tweaks, and project initialization with WARP.md files. Warp Code is leading the Terminal-Bench (52%), outperforming Claude Code by a wide margin. It also climbed to #3 on SWE-bench Verified (75.8%) using GPT-5 high-reasoning models.

Key Highlights:

1. Agent steering with native code review - View and edit agent changes as they're written, with diff views that work against your current branch and allow line-by-line reprompting without switching to GitHub.

2. Project setup and slash commands - Initialize projects with /init to create version-controlled WARP.md files, add global rules and MCP servers, plus define agent profiles with specific models and permissions for different tasks.

3. Proactive error fixing - The coding agent automatically suggests fixes for compiler errors and merge conflicts without being prompted, letting you dismiss, refine, or start conversations from suggestions.

4. Validated performance - Beta users generated over 150 million lines of code weekly with 97% acceptance rates, reporting average daily time savings of one hour, with a quarter saving over two hours.

The growth numbers back up what developers are experiencing firsthand - Warp is adding roughly $1M in annual recurring revenue every 10 days, with revenue up 20x this year. It's clear that more teams are betting on vibe coding with AI agents as the future of coding.

Don’t be the one putting your company at risk 😳

Most meetings include confidential data - from project details to client information.

But not every AI notetaker protects that information with the security your company deserves.

If the AI notetaker you’re using trains their AI models on your team’s conversations, you could be putting your company at risk without realizing it.

This AI Meeting Notetaker Security Checklist helps you avoid that.

In just two minutes, you’ll learn the 7 checks to ensure your team’s AI meeting notes stay private and secure.

Don’t let your meetings become someone else’s dataset.

Get the checklist

Quick Bites

OpenAI released GPT-5 Prompt Optimizer
GPT-5 Prompt Optimizer is a free tool available in the OpenAI Playground that automatically refines your prompts to be clearer and more effective for GPT-5 and other models. It fixes common prompt-writing mistakes by taking your original prompt and creating an optimized version. It's especially useful for migrating prompts from older models, which might have relied on verbose or contradictory instructions, to the more highly steerable GPT-5.

Anthropic has tripled its value in the last 6 months
Anthropic just closed a jaw-dropping $13 billion Series F at a $183 billion post-money valuation - about 3x its valuation from March 2025. The company’s ARR has rocketed from ~$1 billion at the start of 2025 to over $5 billion by August, fueled by 300,000+ business customers and a seven-fold jump in large accounts; Claude Code alone is pulling in $500 million in run-rate with usage up more than 10×. Anthropic now stands among the top four most-valuable private firms globally, banking on enterprise traction, AI safety research, and international expansion.

A USB-C key that turns your phone into an AI assistant
A startup called Dafdef is shipping what might be the first physical AI agent for mobile: a USB-C key that plugs into your iPhone and literally controls your screen - tap, swipe, and type across all your apps - essentially turning your phone into an automated assistant. It's designed for the "watch it work or ignore it entirely" crowd—you can pause, edit, or unplug anytime if the AI goes rogue.

Mistral Le Chat with 20+ custom MCP Connectors
Mistral just turned Le Chat into a proper enterprise hub with 20+ MCP connectors spanning data platforms, productivity tools, and development workflows - think direct integrations with Databricks, GitHub, Notion, and Stripe. The new Memories feature adds persistent context across sessions with granular user control, plus a thoughtful import tool for migrating your ChatGPT conversation history. With everything available on the free tier and flexible deployment options, Le Chat is making a compelling case for switching from ChatGPT and Claude.

Tools of the Trade

1. VibeFlow - Generates full-stack web apps from natural language prompts, showing the backend logic in n8n-style visual, editable workflows. It targets semi-technical users who want AI-generated frontends but need visibility and control over their backend processes.

2. Dedalus Labs - Connects any LLM to any MCP server through a single API endpoint. It hosts and manages your MCP servers, including configs, hosting, scaling, and model hand-offs. No Docker, no YAML.

3. Skipper - A CLI tool that enables CLI agents like Claude Code and OpenAI Codex to access and control your desktop. Do not use it for important work as it lets Claude autonomously take actions on your computer. Try running it in a sandbox or a fresh browser.

4. Sosumi.ai - Translates Apple Developer documentation into Markdown that AI models can actually read, since the original docs require JavaScript rendering. Offers MCP server functionality for automated documentation searches.

5. Awesome LLM Apps: A curated collection of LLM apps with RAG, AI Agents, multi-agent teams, MCP, voice agents, and more. The apps use models from OpenAI, Anthropic, Google, and open-source models like DeepSeek, Qwen, and Llama that you can run locally on your computer.
   (Now accepting GitHub sponsorships)

Hot Takes

1. Ilya Sutskever: bald

   Demis Hassabis: bald

   Noam Shazeer: bald

   Greg Brockman: bald

   
   forget AGI.

   forget curing cancer.

   cure baldness now.

   My hairline is on gradient descent. ~
   Yuchen Jin
   
   

2. unpopular opinion: offering AI compute for free is bad for the industry, and players like Google should stop doing it.

   It creates an unrealistic expectation that the models are actually cheap to run, and causes people to overuse compute on worthless tasks ~
   mark

That’s all for today! See you tomorrow with more such AI-filled content.

Don’t forget to share this newsletter on your social channels and tag Unwind AI to support us!

Unwind AI - X | LinkedIn | Threads

Awesome LLM Apps | Sponsor Us

PS: We curate this AI newsletter every day for FREE, your support is what keeps us going. If you find value in what you read, share it with at least one, two (or 20) of your friends 😉