---
title: "MCP and Make Power-Ups, meeting booking and a Search API"
canonical_url: "https://chatthing.ai/blog/mcp-servers-make-scenarios-and-meeting-booking"
last_updated: "2026-10-27T14:03:15.937Z"
meta:
description: "Give your AI agent tools from any MCP server, let it run your Make scenarios and book Cal.com or Calendly meetings without leaving the chat."
"og:description": "Give your AI agent tools from any MCP server, let it run your Make scenarios and book Cal.com or Calendly meetings without leaving the chat."
"og:title": "MCP servers, Make scenarios and meeting booking - Blog - Chat Thing"
"twitter:description": "Give your AI agent tools from any MCP server, let it run your Make scenarios and book Cal.com or Calendly meetings without leaving the chat."
"twitter:title": "MCP servers, Make scenarios and meeting booking - Blog - Chat Thing"
---
[← Back to blog](https://chatthing.ai/blog)**Changelog**# **MCP servers, Make scenarios and meeting booking**
~ 10 min read
---
---
### **On this page**
1. [Give your agent tools from any MCP server](#give-your-agent-tools-from-any-mcp-server)
1. [Connect once, use it in any agent](#connect-once-use-it-in-any-agent)
2. [You decide exactly what it can do](#you-decide-exactly-what-it-can-do)
3. [Tools stay up to date](#tools-stay-up-to-date)
4. [Not to be confused with the Chat Thing MCP server](#not-to-be-confused-with-the-chat-thing-mcp-server)
2. [Run your Make scenarios from the chat](#run-your-make-scenarios-from-the-chat)
3. [Book meetings without leaving the chat](#book-meetings-without-leaving-the-chat)
1. Cal.com: your agent books it
2. [Calendly: your form, inside the chat](#calendly-your-form-inside-the-chat)
4. [Search your agent's knowledge from your own code](#search-your-agents-knowledge-from-your-own-code)
5. [Faster, safer Notion syncs](#faster-safer-notion-syncs)
6. Claude Sonnet 5.5
7. [Smaller improvements](#smaller-improvements)
1. [Add voice minutes now](#add-voice-minutes-now)
2. [Daily message limit per visitor](#daily-message-limit-per-visitor)
3. [Also](#also)
8. [Try it](#try-it)
You've always been able to wire your Chat Thing agent into your other systems with the **API request** Power-Up. It works, but every action means setting up an endpoint, its inputs and its authentication by hand.
This release makes that a few clicks. If a tool publishes an MCP server, connect the whole server and choose which of its tools your agent gets. If you've already built an automation in Make, your agent can run it as it is. And visitors can now book a meeting without leaving the chat. Developers also get a Search API, Notion syncs are much quicker, and Claude Sonnet 5.5 joins the model line-up.
## Give your agent tools from any MCP server
Lots of tools now publish an MCP server: a standard way for AI agents to read and act on what's inside them. Linear, Notion and HubSpot all have one, and so might the systems your own developers have built. The **MCP server** Power-Up lets your agent use those tools in the conversation. Unlike the API request Power-Up, where you set up each call yourself, one connection gives your agent every tool the server offers, with names, descriptions and inputs already filled in.
### Connect once, use it in any agent
You connect a server once for your team, then choose which of its tools each agent can use.
1. Open your agent, go to **Power-ups** and click **New power-up**.
2. Choose **MCP server**.
3. Paste the server's URL and click **Connect**. If the server uses OAuth, a window opens so you can sign in and approve access. If it takes an API key, add it under **Authentication**.
4. Tick the tools this agent can use, then click **Save**.
Every connected server is listed on the **Connections** page, along with the Power-Ups that use it. Sign-ins renew automatically, and if a server ever rejects Chat Thing's access, team owners and admins get an email with a link to reconnect.
### You decide exactly what it can do
Connecting a server doesn't hand your agent everything on it.
- **Every tool starts switched off.** You tick the ones the agent can use. Tools the server adds later stay off until you turn them on.
- **Read-only and Destructive labels** show what the server says each tool does, and **Enable all read-only** ticks the safe ones in one go.
- **Fixed arguments** lock an input to a value the agent can't see or change. Use them to keep a search to one project, or one customer account.
- **Your own descriptions** replace the server's when you want to tell your agent more precisely when to use a tool.
Your agent uses the server with the access you granted, for every visitor. So connect servers you trust, and only switch on tools you're happy for any visitor to trigger.
### Tools stay up to date
Chat Thing reads the tool list from the server while your agent chats, refreshing every few minutes. When a tool's description or inputs change on the server, your agent picks that up without you saving anything. The tool picker also shows how many tokens your selection adds to each message, so you can keep the agent focused on the tools it actually needs.
The MCP server Power-Up is available on Basic, Standard, Pro and Enterprise. It works with remote servers that use the Streamable HTTP transport. Servers that only run on your own computer can't be connected. Everything is covered in [Give your agent tools from an MCP server](https://chatthing.ai/docs/power-ups/mcp-server).
### Not to be confused with the Chat Thing MCP server
In June we launched the [Chat Thing MCP server](https://chatthing.ai/blog/chat-thing-mcp-server), which lets Claude, ChatGPT and Cursor build and manage your agents. This Power-Up works the other way round: your agent uses other people's MCP servers.
The two now meet in the middle. If you build with Claude or ChatGPT, you can connect a server and set up the Power-Up from there too. We've also given every Chat Thing MCP tool a clear title and marked which ones only read and which change something. Your assistant should stop asking you to confirm harmless lookups, and keep asking before anything that changes your agents.
## Run your Make scenarios from the chat
If your team already automates things in [Make](https://www.make.com), your agent can now start those automations itself. The new **Make scenario** Power-Up turns one of your on-demand scenarios into something your agent can run during a conversation: add a lead to your CRM, open a support ticket, look up an order.
- **The scenario's inputs become what your agent fills in.** Each input has a description that tells the agent what it's for, and you can edit it. Give an input a fixed value, such as the source of a lead, or tell the agent to ignore it.
- **Results come back to the agent.** End the scenario with a **Return output** module and your agent answers with whatever it returns. Make waits up to 40 seconds for a run. If one takes longer, your agent tells the visitor it's still running.
- **Mistakes get a second try.** If Make rejects an input, your agent sees Make's reason and can fix it or ask the visitor again.
- **Connect your Make account once.** Add your API token once for your team and every Make Power-Up, on any agent, can use it. The token is stored encrypted and never shown again, and you replace it in one place from **Connections**.
Add one Power-Up for each scenario you want your agent to run. It works with every Make plan, including Free, and the Power-Up is available on Basic, Standard, Pro and Enterprise. See [Run Make scenarios from your agent](https://chatthing.ai/docs/power-ups/make).
**MCP or Make?** If the tool you want already has an MCP server, the MCP Power-Up gives your agent its tools directly. If it doesn't, or you want to join several apps together in one step, build it as a Make scenario. The [API request](https://chatthing.ai/docs/power-ups/call-your-api) and [Zapier](https://chatthing.ai/docs/power-ups/zapier) Power-Ups are still there for everything else.
## Book meetings without leaving the chat
We added these two Power-Ups in September and never properly told you about them, so here they are.
"Can we book a call?" is one of the best things a visitor can ask. Answering it with a link to a separate booking page is how you lose them. Now your agent can take the booking in the conversation.
### Cal.com: your agent books it
Add the **Book a meeting (Cal.com)** Power-Up and paste your event link. Public events need no API key and no account connection. Click **Check event** and Chat Thing shows the event's next free times, so you know it works before a visitor tries it.
When someone asks for a meeting, your agent reads your real availability and offers it. In your website chat, that's a calendar in your agent's colours, in the visitor's own time zone. If they've already given their name and email, those are filled in. Once they confirm, the agent books it and shows the meeting link, cancel and reschedule links, and buttons to add it to their calendar. Visitors can ask the agent to cancel later, too.
On Slack, WhatsApp, Telegram and email, there's no calendar to click, so your agent lists the times as text and books whichever one the visitor picks. Prefer your own booking page? Switch **Booking form** to **Cal.com's own booking form** and it opens inside the chat instead. See [Book meetings with Cal.com](https://chatthing.ai/docs/power-ups/book-meetings-cal-com).
### Calendly: your form, inside the chat
Add the **Book a meeting (Calendly)** Power-Up, paste your event link and save. There's no OAuth, no API token and nothing to install in Calendly. When a visitor wants a meeting, your agent opens your Calendly booking form in the conversation with their name and email already filled in.
Calendly keeps its scheduling API behind a paid add-on, so this one does less: your agent can't read your availability, and cancelling or rescheduling happens through Calendly's confirmation email. If you want the agent to offer times and book them itself, use Cal.com. See [Book meetings with Calendly](https://chatthing.ai/docs/power-ups/book-meetings-calendly).
Both booking Power-Ups are available on Basic, Standard, Pro and Enterprise.
## Search your agent's knowledge from your own code
Sometimes you want your agent's knowledge without the agent. A search box in your own help centre. An internal tool that needs the right page from your docs. Or an integration that has to check what each user is allowed to see before anything reaches the agent.
The new [Search API](https://chatthing.ai/docs/developers/search-api) runs the same search your agent uses when it answers. Send a question and get back the most relevant passages, best match first. Each passage comes with the data source and row it came from, such as a Notion page, a web page or a file, so you can match it to your own records.
- Narrow a search to one data source, or set the minimum relevance yourself.
- It covers every type of data source on the agent, not only the ones you manage through the API.
- Each channel can search 120 times a minute, and searches don't use message tokens.
It's part of the API channel, which is on the Standard, Pro and Enterprise plans.
## Faster, safer Notion syncs
If your agent learns from Notion, its scheduled syncs now only fetch the pages that have been edited since the last one. On a 450-page workspace with nothing changed, a scheduled sync went from about 17 minutes to about half a minute in our testing. A sync you start yourself with **Synchronise** still fetches every page.
- **A Notion outage no longer costs you pages.** If Notion is down or rate-limits a sync, the sync stops and leaves your agent's Notion content as it was.
- **Databases inside your pages are included.** Each entry is synced as its own item, so your agent can answer from them.
- **Your agent knows which page it's reading.** Every passage from Notion now carries its page title and where it sits in your workspace, not just the first passage on each page. Pages you've published to the web also carry their public link.
See [Add Notion pages](https://chatthing.ai/docs/knowledge/notion) for how long syncs take and what counts as an edit.
## Claude Sonnet 5.5
Anthropic's Claude Sonnet 5.5 is now in [Chat Thing's model line-up](https://chatthing.ai/models).
| Model | Provider | Context window | Message tokens |
| --- | --- | --- | --- |
| Claude Sonnet 5.5 | Anthropic | 1,000,000 | Heavy |
It costs the same as Claude Sonnet 5 in message tokens, supports vision and Power-Ups, and always does a little reasoning before it answers. If your agent already runs on Sonnet 5, Sonnet 5.5 is the obvious one to try next. Test it with your own questions before you switch a live agent: check the answers, the tone and whether it calls the right Power-Ups. Choosing a model is available on Standard, Pro and Enterprise.
## Smaller improvements
### Add voice minutes now
Running low before your renewal? Go to **Billing**, find **Voice calls** and click **Add minutes now**. The minutes join your current allowance as soon as the payment goes through, and you can add the same packs to every month. Packs added this way are charged at the full pack price and expire with the rest of this period's minutes. See [Voice minutes and phone numbers](https://chatthing.ai/docs/account/voice-minutes#add-minutes-now).
### Daily message limit per visitor
Cap how many messages one visitor can send your website agent in a day. Turn on **Daily message limit per visitor** under **Usage limits** in your agent's settings. Visitors are identified by their network address, so people sharing a network share the limit, and your own team is never limited.
### Also
- **Team invitations hold on.** An invite now survives signing up, signing in with Google or GitHub, and confirming your email in another tab, and email addresses match whatever their capitalisation.
- **Our DPA covers UK GDPR.** The [Data Processing Agreement](https://chatthing.ai/legal/dpa) now puts UK law first and covers UK data transfers, for anyone who needs it for their own compliance paperwork.
## Try it
Pick one question your agent can't answer today because the answer lives in another tool. If that tool has an MCP server, add the **MCP server** Power-Up and switch on just the read-only tools that answer it. If it doesn't, build a small on-demand scenario in Make and add the **Make scenario** Power-Up. Then ask your agent in the test chat. And if visitors ever ask to talk to you, add a booking Power-Up while you're there.
### **Related articles **
[](https://chatthing.ai/blog/citations-call-summaries-and-eight-new-models) [Inline citations, phone call summaries and eight new models
](https://chatthing.ai/blog/citations-call-summaries-and-eight-new-models)
Chat Thing agents now cite only the pages an answer used, phone agents send a summary after every call, and eight new models join the line-up.
[](https://chatthing.ai/blog/voice-calls-and-phone-numbers) [Chat Thing agents can talk: voice calls and phone numbers are here
](https://chatthing.ai/blog/voice-calls-and-phone-numbers)
Chat Thing agents can now hold voice conversations in your web chat and answer a US or UK phone number, with the same knowledge, power-ups and AI model you've already set up.
[](https://chatthing.ai/blog/glm-5-3-flash-supportbench-and-link-controls) [GLM 5.3 Flash lands in Chat Thing, plus SupportBench and link controls
](https://chatthing.ai/blog/glm-5-3-flash-supportbench-and-link-controls)
GLM 5.3 Flash is now available in Chat Thing, alongside SupportBench model testing and control over where links in embedded agents open.
- **Previous **
- [← Introducing voice calls: your AI agent, now speaking](https://chatthing.ai/blog/introducing-voice-calls)