11 posts tagged with "Features"

TL;DR​

• You can attach an A2A agent to any MCP server in your workspace—the same catalog surface you use from Claude or Cursor—not only from the Hub.
• Agent instructions are enough to run manually or from chat; a heartbeat checklist is optional and mainly for recurring scheduled runs.
• On a bundle MCP connection, agent tools use the same names as the Hub (get_agents, upsert_agent, trigger_agent_run, and the rest).

It's Tuesday standup and someone asks whether yesterday's Google Ads spend spiked for a client you do not have open in another tab. You already have the MCP server in Cursor. What you want is a named assistant on that server that you can wake on demand or on a clock.

That is the workflow we optimized this week.

MCP solved the first half of the problem for us: connect ChatGPT, Claude, or Cursor to real software, with credentials scoped properly and tools grouped into servers you can actually manage.

The second half took longer to name. Who runs the assistant when nobody is sitting in chat?

The next shift in AI tools is not only smarter text. It is AI you can see — maps, layouts, and lists that show up inside the conversation when a paragraph is not enough.

MCP Apps (the official extension for interactive UI in MCP hosts) are still early. Iframes feel clunky on some hosts, CSP trips up builders, and Claude and ChatGPT do not behave identically. When it works, though, the conversation changes: you get something you can click, drag, or approve — not just read.

Below are three real screenshots from ChatGPT — Canva, Tripadvisor, and Charge Atlas. No stock art, no protocol demo tiles.

MCP Apps are interactive tool outputs that render as rich HTML and CSS inside supported AI clients. Instead of stopping at plain text or raw JSON, an MCP tool can return a full, styled experience: dashboards, charts, forms, data tables, and cards that users can read and interact with in context.

The Model Context Protocol already made it possible for assistants to call tools against live systems. MCP Apps extend that idea: they turn those calls into something that feels closer to a small application than a chat transcript. For anyone searching mcp app or mcp apps, this guide explains what that means technically, why teams care, and how to build one.

Traditional MCP tools return text. AI clients display text well. Many real workflows still need visual output — a stock dashboard, a compliance summary, a project status board, a customer profile. MCP Apps bridge the gap between “the model called a tool” and “the user got something they can actually use on screen.”

Today, MCP Apps are live as the first official MCP extension. Tools can now return interactive UI components that render directly in the conversation: dashboards, forms, visualizations, multi-step workflows, and more. It is ready for production.

Tool overload is real.

It shows up as lag. Wrong tool picks. Weird, half-finished workflows. Or the model just dumps a wall of raw data at you and calls it a day.

We’ve always had a simple answer: keep bundles focused. 5–15 tools for one job.

That still works great.

But sometimes you do want a big bundle. A real “everything I use for this role” bundle.

Now you can do that without turning your AI into a confused mess.

Every bundle can run in Dynamic.

Your AI assistant calls a tool through MCP. What actually happened? Did it work? How long did it take? What did it send, and what did it get back?

Before now, you'd dig through logs or just hope your AI would explain what happened. That's not enough.

We built complete tool execution tracking right into MCP Bundles. Every tool call gets logged with full context—timing, credentials used, results, everything. You can see exactly what your AI is doing, debug failures fast, and understand your tool usage patterns.

Tool overload is real. Give AI 50 tools and it gets confused—slow, wrong tool selections, data dumps instead of answers. We've always solved this with focused bundles: give AI 5-15 tools for a specific workflow, and it works great.

But what about when you need data from multiple services at once?

That's why we built the Hub. It uses programmatic tool calling—AI discovers tools on-demand and writes code to orchestrate them—so you can work across all your connected services without the overload problem.

This builds on recent research from Anthropic—their work on advanced tool use and code execution with MCP. We took these patterns and made them accessible to anyone with an MCPBundles account.

Ask your AI about pipeline data and get a real dashboard. Not text. Not a description. An actual interactive visualization you can click, filter, and explore.

If you've ever installed an app on your computer—double-clicking a .dmg file on Mac or a .exe on Windows—you already understand .mcpb files.

Anthropic introduced the .mcpb extension (MCP Bundle) as the standard packaging format for distributing MCP servers. Think of it as the "app bundle" for AI tools: one file that contains everything needed to give your AI assistant new capabilities.

If you've used AI assistants, you've probably hit this frustration: you need to re-explain context every time you start a new conversation. That database schema you carefully described yesterday? Gone. The code snippet you asked about last week? You're pasting it in again.

MCP Resources solve this. They let you give your AI access to files and data that persist across conversations — like giving your assistant a file cabinet instead of asking them to memorize everything.