# Top MCP's You Should be Using for Claude, ChatGPT and Gemini

An episode of [AI Chat: AI News on Claude, ChatGPT, Gemini, Grok, Perplexity](https://invisiblenetwork.ai/podcast/1669990482), 18 Jun 2026, 17 min.

Audio: https://content.rss.com/episodes/365073/2926110/ai-chat-podcast/2026_06_18_21_07_27_aa326fcd-6f6d-4bb7-b50b-03b95397fff8.mp3

Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode! In this episode, we explore the functionalities of MCPs and how they enhance the capabilities of AI tools like Claude and ChatGPT. We also discuss the differences between MCPs and APIs, share practical use cases, and highlight some of the most effective MCPs available for maximizing your AI integrations. Chapters 00:00 Introduction to MCPs 02:00 Understanding APIs vs. MCPs 03:59 Setting Up MCPs Easily 09:58 Top MCP Tools and Recommendations 15:01 Benefits of Using MCPs Show Links • Get the AI Box MCP: https://aibox.ai/mcp • How I Grow and Scale My Business with AI: https://www.skool.com/aihustle • Get the AI Chat Daily Newsletter: https://www.aichatdaily.com/newsletter

## Transcript

[0:02] Welcome to the AI Chat Podcast. I'm your host,
[0:04] Jaden Schaefer. Every day I cover cutting -edge
[0:07] AI news and talk with the leaders behind it,
[0:09] breaking down what it means for your life and
[0:11] business. Welcome to the podcast. Just yesterday,
[0:17] I released an episode where I said that I have
[0:19] created a tool, an MCP with my company, AI Box,
[0:22] that gives Claude the ability to generate images.
[0:25] It gives ChatGPT the ability to generate video
[0:28] once Sora dies and they're going to have to use
[0:30] Google VO3. And it gives something like Gemini
[0:33] the ability to generate audio with 11 Labs. So
[0:36] you essentially can take any of the main AI tools
[0:39] you use day to day. And I think Claude, ClaudeCode,
[0:41] ClaudeCode is kind of one of the most popular.
[0:44] AI models inside of it. Now, I had a lot of people
[0:47] asking me number one, how MCPs work, what MCPs
[0:50] are, what the difference between MCP and API
[0:52] is. So I want to make an episode where I'm going
[0:54] to say some of the cool functionality that I've
[0:56] created with AI box and some of the cool things
[0:58] I've learned on the journey there. But I'm also
[1:00] going to break down for anyone curious what the
[1:02] top MCPs are that you should be using with cloud
[1:05] or whatever AI tools and assistance you're building
[1:07] with, and how these MCPs and APIs actually work,
[1:11] what the difference is between them, and how
[1:12] you can get the most out of them. As just two
[1:15] seconds of background, I'm sure most of you know,
[1:16] but just to define the terms and explain what
[1:19] an API is, it's basically the way for one piece
[1:21] of software to talk to another piece of software.
[1:23] And the difference between an API and an MCP
[1:26] is that an MCP is basically a standard way for
[1:29] an AI assistant, you can think of something like
[1:31] Claude, to basically discover and use tools,
[1:35] any apps, any files that are inside of a database
[1:37] or inside of a service. So an API, a lot of times,
[1:40] not always, it's kind of like, I guess a good
[1:43] analogy for this. is kind of a universal adapter,
[1:46] right? So every country in the world has a different
[1:47] outlet plug. That's kind of like an API. It's
[1:50] kind of like a very specific way to use their
[1:53] software. And an MCP is more like a universal
[1:56] adapter that will fit into any of those plugs
[1:58] because the way an MCP works is it's basically
[2:01] kind of like a documents page. It's basically
[2:04] an explanation to something like Claude of all
[2:06] of the ways that your API or all of the ways
[2:09] your software operates and how you can interact
[2:11] with it versus like, API was like very structured,
[2:14] very rigid. And as an example from something
[2:16] I've built, I you know, I recently built a tool
[2:18] that helped me with posting some of my podcast
[2:21] episodes. But I have three different podcast
[2:23] distribution companies that I use, I use rss
[2:26] .com, I use a cast, I use art 19. All of them
[2:28] have API's that helped me with my podcast publishing.
[2:31] So a way for me to, you know, upload an episode
[2:34] to my platform and have it go and publish it
[2:36] onto one of those distribution networks. But
[2:38] the problem is every single one of those API's
[2:41] like the way they work. the way that you integrate
[2:43] them into software is different. So what all
[2:45] of those companies do and what everyone has done
[2:47] with APIs, they've always kind of published a
[2:50] documents page that explains how you structure
[2:53] your API calls, how you have to give them code,
[2:56] how they'll give you code back. And if you have
[2:57] anything wrong in that code that you send them,
[2:59] you get back, it crashes and breaks and the API
[3:01] doesn't work. So it's been pretty, you know,
[3:03] it's frustrating. And a lot of times when I was
[3:05] first building with tools with Claude and building,
[3:08] you know, that particularly that use case, that
[3:10] podcast distribution, tool that I created, I
[3:12] would literally say, Hey, Claude, here's my API,
[3:15] which I'll tell you why that's a bad thing in
[3:17] the future. But I'd say, here's my API key. And
[3:20] here is a link to their documents page, read
[3:22] this entire page of context pages and pages to
[3:25] figure out how the heck their API works to integrate
[3:27] it. And it still wouldn't work all the time,
[3:28] we'd have to do tons of tests, sometimes I have
[3:30] to get on with their customer support, and then
[3:31] have to look at what I was doing and looking
[3:33] at my code errors. And Claude was telling me
[3:35] stuff. And anyways, nightmare so then we have
[3:37] a beautiful you know solution which is an mcp
[3:40] which is basically they give you a link that
[3:43] you can give claude that thoroughly helps it
[3:45] understand how your api works and newsflash mcp
[3:49] is also leveraging an api a lot of the time but
[3:51] it's basically the universal adapter to explain
[3:54] to the ai agent how to successfully use these
[3:56] apis and it's a standardized way to do it so
[3:58] So now instead of struggling with an API and
[4:01] trying to figure out how to work with a software,
[4:02] you just give the MCP link to your AI agent and
[4:06] it will figure it all out for you. You don't
[4:08] have to finick with anything anymore. And it's
[4:10] really easy to set up an MCP, by the way, in
[4:12] Cloud. And I didn't really realize this. I think
[4:15] a lot of us have probably done this before with
[4:16] Cloud or with ChatGPT where you go to the connectors
[4:19] panel, right? And it shows you like, hey, look,
[4:21] here's like 20 different softwares that integrate
[4:23] with Cloud or with ChatGPT or Gemini. And I'll
[4:25] be honest, like when I first saw all of the list
[4:27] of connectors, on Claude, I was kind of disappointed.
[4:29] I'm like, oh yeah, there's 20 things here. I
[4:31] might use like three of them, like Gmail and
[4:33] Slack and Google Drive, but like, I don't really
[4:34] use all of these other random things. And there's
[4:36] tons of, you know, software that I like and I
[4:38] use, but it wasn't listed on there, right? Like
[4:40] for example, my email newsletter software, Beehive,
[4:43] like it's just not, you know, they didn't have
[4:45] that. And so Beehive actually went and created
[4:47] their very own MCP and it's a link you can get,
[4:50] but it's not inside of the Claude Connections
[4:52] kind of like a store or whatever. So what you
[4:55] can do, it's actually super easy. You grab that
[4:57] link from Beehive. You click add your own connection.
[5:00] You paste that link in and it's going to have
[5:02] a pop up come up on your screen that asks you
[5:04] to log in to your account on Beehive or you could
[5:08] do this also with AI box. So you put the AI box
[5:10] MCP link in as a pop up that says log into your
[5:13] AI box account. You log in just like it like
[5:15] a login with Google kind of thing. And all of
[5:17] a sudden that is how it authenticates. And now
[5:19] it knows that it has your login credentials and
[5:22] it's using those and it has this MCP link. So
[5:25] understands how to interact. your software and
[5:27] you're off to the races. At this point, you don't
[5:28] need to do anything. You don't need to mess around
[5:30] with APIs or figure out how anything works. You
[5:31] can just tell Claude, hey, go to Beehive and,
[5:34] you know, help me figure out my analytics for
[5:36] my newsletter that I've been publishing or go
[5:38] to AI Box and build a tool that can help me automate
[5:40] generating thumbnails much faster. Just a PSA.
[5:44] If you use Claude, Claude Cowork or Claude Code,
[5:47] and you'd like to be able to work with all of
[5:50] the tools that you have, there's probably MCPs
[5:52] for all of the different software you use. You
[5:54] just have to go and You just have to go Google,
[5:57] you know, like Vercel MCP URL and you go get
[6:00] that and manually add it. So anyways, hope this
[6:03] was useful. Honestly, it has 10x what I've been
[6:06] capable of doing with a cloud. And so in the
[6:08] case of AI box, what you can do with it, you
[6:10] go and add the AI box MCP URL, you go and authenticate
[6:13] and log in, and then you're able to. Anytime
[6:17] you build like a tool on AI box, if you've built
[6:20] like an image generator or like a newsletter
[6:21] generator or, you know, some sort of like audio
[6:25] or video generation, you know, creating ad assets
[6:27] like a tool, you know, you can build those kind
[6:29] of no code tools on AI box with all the different
[6:31] AI models. If you've created something like that,
[6:33] you can go to Claude and say, hey, Claude, I
[6:35] need to go and generate like 10 new images for
[6:38] my Facebook ad campaign for my product XYZ. Go
[6:42] and use my, you know, my. my box over on AI box
[6:46] and you say the name of the box to generate 10
[6:49] new things and make sure that they have like
[6:50] this new direction or this new style or theme
[6:54] that we're thinking of or this new product. And
[6:56] it will go and access that box on AI box because
[6:59] it has the MCP connection. It will generate your
[7:01] assets and then it will let you know and give
[7:03] you links back to your assets where you can go
[7:05] download the images or videos or whatever you've
[7:07] created. Or if it's text, I think we might even
[7:09] port it straight into your chat message so you
[7:11] can see it right there. So that's how it works
[7:13] with AI box. I'm really super excited about it.
[7:16] And by the way, if you haven't tried out something
[7:19] like this, MCPs are really useful, specifically
[7:21] if you have to do generations that have many,
[7:25] many different iterations. And so an example
[7:28] of this. I was recently creating a Bible study
[7:31] app and I wanted it so you could tap on any verse
[7:33] in the Bible and get it to, you know, like, give
[7:36] me give me the background, give me the explanation,
[7:37] give me the context of this verse, like explain
[7:39] this verse to me. I feel like very often I'm
[7:41] reading something. I have no idea what it's saying.
[7:43] Right. And so what I did is I had Claude go and
[7:45] generate an explanation for every single verse
[7:47] in the Bible. Now, the problem with that is there
[7:49] is like. 30 to 45 ,000 verses in the Bible, in
[7:54] the Old and New Testament. It is insane. And
[7:56] generating that takes a very long time. And I
[7:58] wanted these to be really high quality generations.
[8:00] I talked about like the historical context and
[8:02] like all this stuff, right? So I had this massive
[8:04] project. I wanted Claude to do it. And the problem
[8:06] was, I would say like, okay, go do these generations.
[8:08] And it'd be like, okay, it's going to take me
[8:10] like two days to do them. And I'm like, okay,
[8:11] you know, I just have to keep like continuing,
[8:13] continue, continue every like couple hours when
[8:15] it does a new batch. The problem was sometimes
[8:17] my computer would get shut off. off. Sometimes
[8:20] the internet would cut out. Sometimes Claude
[8:22] would just, you know, like their connection would
[8:24] sever, their servers would sever. Sometimes I'd
[8:26] hit a rate limit, whatever. There's a hundred
[8:28] different reasons why halfway through a generation
[8:30] of a batch, it would just crash. And what happens
[8:33] when it would crash is I would basically lose
[8:35] the entire generation. So it could be running
[8:36] for an hour, crash at minute 50, and I would
[8:39] lose like half or all of everything it was generating.
[8:44] It was very frustrating. So what I'm able to
[8:47] do is go create a box on AI box where it's just
[8:50] like the MD file that you're giving Claude to
[8:52] do the project. And in fact, you can take that
[8:53] MD file and bring it over to AI box and just
[8:55] give it a prompt in the vibe builder to create
[8:57] that tool. But I created a box in AI box that
[9:00] can generate these verse descriptions. And then
[9:03] over on Claude, I say, hey, look, I need to go
[9:06] generate like 30 ,000 of these verse descriptions.
[9:09] Use this particular box to do it. And a couple
[9:12] of things, it's capable of doing a couple of
[9:13] things. Number one, because it's running the
[9:16] processing on AI box, you can instead of having
[9:18] to wait two days for it to generate everything,
[9:20] it technically could open, you know, 100 or 1000
[9:23] or, you know, 20 ,000 instances of AI box and
[9:26] get to generate all of them in parallel at once.
[9:29] So number one, that saves you a ton of time rather
[9:31] than having to wait for Claude to do it one by
[9:33] one, you could just have AI box, you know, do
[9:35] a million of them in parallel, not literally
[9:38] a million, but a ton of them in parallel. And
[9:40] number two, if any of them crashes, or particularly
[9:43] if Claude crashes in the middle of any of these
[9:46] generations... AI box still has all of your assets
[9:49] saved over there. So, you know, let's say like
[9:52] cloud crash and it's like, oh shoot, I crashed.
[9:53] You're like, okay, great. Well, AI box has them
[9:55] all generated. There's a documents folder. There's
[9:59] a file section over on AI box. You can actually
[10:02] go and access everything that has been created
[10:04] on any boxes or tools on AI box. So nothing is
[10:07] a lost in the case of a lot of my generations.
[10:10] And it's not the end of the world when it's,
[10:11] you know, trying to generate like some piece
[10:13] of text, but I was doing something very similar
[10:14] with 11. where I was generating like a whole
[10:18] bunch of sleep history podcast episodes. It's
[10:20] a new thing I was trying. Someone in my studio
[10:22] wanted a new project. So we generated a bunch
[10:25] of sleep history podcast episodes. They're like
[10:27] two to four hours long. And I had a ton of free
[10:31] credits from 11 labs that they were giving me.
[10:33] And they were worth thousands of dollars. And
[10:37] so I spent about $1 ,200. unbeknownst to me,
[10:42] because I told like my my AI generation tool
[10:45] to like go generate these podcasts, put them
[10:48] in a specific folder, I was using cloud. And
[10:50] it crashed at some point in that generation,
[10:53] I had like an API, but it all crashed. And I
[10:56] said, you know, what's the status of it? And
[10:57] it's like, Oh, shoot, yeah, we crashed generating
[10:59] all of these things, your your 11 labs API crashed.
[11:01] And so we actually don't have anything saved.
[11:04] And I'm like, Okay, well, How many, like, had
[11:06] we generated? I think it had got stuck on a loop
[11:08] or a glitch. Something happened. And it's like,
[11:10] okay, well, we actually spent about $1 ,200 of
[11:13] credits on 11 labs. But yeah, it all crashed
[11:16] and we don't have any of that. So that was a
[11:18] very painful experience to just lose $1 ,200
[11:21] for no reason. I mean, for the crashes and it
[11:24] was unrecoverable. If you had done that with
[11:26] AI Box, right? So created a box on AI Box that
[11:29] had an MD file, had an explanation of the type
[11:32] of... you know, the sleep history podcast that
[11:34] I was going to create. And you, all you have
[11:36] to do is just give it the title of the episode
[11:37] and then it will write the whole episode and
[11:38] generate it and give you your audio file. If
[11:40] you'd done that with AI box, then there wouldn't
[11:43] be the crashing. And if Claude crashed for any
[11:46] reason, AI box would have safely generated and
[11:48] be able to get you your asset. So instead of
[11:50] me saying, Oh shoot, I just lost $1 ,200 of 11
[11:54] labs generations. They would all, even if Claude
[11:56] crashes, they would all still be safe on AI box.
[11:58] So this is what we're trying to build with the
[12:00] MCP. We're basically trying to make it your pipelines
[12:02] of AI. content much more reliable. And at the
[12:05] same time, we're trying to make it so that if
[12:08] any crashes happen on Claude or on any of the
[12:11] software that you're vibe coding, if you have
[12:13] the engine of it being run on AI Box, you never
[12:16] lose your assets. You never lose access to your
[12:19] assets. And of course, the other thing that people
[12:21] love about it is if you're creating tools that
[12:23] generate images, audio, video, Claude just can't
[12:26] do that. It doesn't have the capabilities. And
[12:28] so if you create a box on AI Box that can generate
[12:31] any of those assets, you can tell Claude to...
[12:33] to call it and to do it. And essentially, you're
[12:36] able to generate, you know, images with Claude,
[12:39] which was never a capability that Claude had
[12:41] before. And it saves you a ton of time for, you
[12:43] know, whether that's generating images for, you
[12:45] know, graphics for your apps or for, you know,
[12:48] blog post titles, like all if you look at AI
[12:51] chat daily .com, my AI news site, all of the
[12:55] title, all of the banner images, and all of the
[12:57] articles are AI generated. And using something
[13:00] like AI box to generate them helps it to be super
[13:02] consistent. and then we don't have that crashing
[13:04] issue. Okay, and of course, also we give chat
[13:07] or cloud the ability to generate images, which
[13:09] it didn't even have before. Okay, I want to talk
[13:11] about some of the top most used MCP tools that
[13:14] are out there. There's really interesting, there's
[13:16] some interesting leaderboards, MCP leaderboards,
[13:18] which are great places to see what ones are actually
[13:21] very useful. Sometimes a company will come out
[13:23] with an MCP, but it doesn't have all of the capabilities
[13:26] of their platform. And, you know, I hate to,
[13:29] I hate to bag on like beehive because I love
[13:31] that platform for email. sending, but their MCP
[13:33] only gives you data about the platform. It doesn't
[13:35] actually help you post to the platform yet. And
[13:37] they said like they're working on it. So there's
[13:39] some, you know, that sometimes the functionality
[13:40] isn't completely there of everything you could
[13:42] do if you were just on the website. But I did
[13:44] want to say some of my favorite MCPs I use and
[13:46] talk about some of the other more popular ones.
[13:48] The number one is that I'm using is the GitHub
[13:50] MCP. And that can also, I think it's like built
[13:53] into Claude. So you probably see that one. You
[13:55] probably don't have to go get a custom connection
[13:56] for it, but I have everything. If you're going
[13:59] to be developing anything, I recommend getting
[14:01] something. like Vercel or Netlify MCP and getting
[14:04] Supabase and then getting GitHub and connecting
[14:08] that stack. You can essentially log into all
[14:10] your accounts and anything you want to build,
[14:12] Claude can go and set up your database. It can
[14:14] set up your hosting and it can set up, you know,
[14:16] where you store your code, which is GitHub and
[14:18] what you're working through. So just those three
[14:20] are basically the most popular ones that I personally
[14:22] am using. I love some of the Google connection
[14:24] ones that are able to do something like Gmail,
[14:27] Google Calendar and Google Drive and interact
[14:29] with all of those for me. those are also super
[14:31] popular. And then a couple that I was seeing
[14:33] people recommend online that are really useful
[14:36] are one of them being Brave Search. It's I think
[14:39] one of the most popular web search MCPs. Basically,
[14:42] the AI can get past any sort of training cutoff
[14:44] and can find current library releases or CVEs.
[14:47] Figma has one, which is interesting when it comes
[14:49] to working between design and code. Canva has
[14:52] some version of this. And then I think Notion,
[14:55] a lot of people use Notion. If you have that,
[14:57] it's a great way to connect AI to some of your
[14:59] workspaces. for syncing your knowledge bases,
[15:01] your documents, and a bunch of your tasks. Now
[15:03] there's tons of places that you can plug this
[15:06] into. You can plug this all into Cursor, Cloud
[15:08] Desktop, Gemini, I think Grok, and a bunch of
[15:12] others. So lots of different places to give these
[15:14] MCPs to and lots of different software. But honestly,
[15:17] chances are, if you have software that you use
[15:19] specifically in your industry and it's not some
[15:21] super niche software, and even honestly, a lot
[15:24] of super niche software even too, it will have
[15:26] an MCP built out for it. So go check, just Google.
[15:29] mcp for your software see if it's there and go
[15:31] and attach it and basically you'll you know save
[15:34] yourself if you're using something like cloud
[15:35] co -work like yes you can get around this and
[15:37] i do this all the time where like it'll say hey
[15:39] i don't have like the even sometimes i have the
[15:41] mcp for example to like super base my database
[15:43] and it's like hey i don't have like the permission
[15:45] to edit this one particular thing can you get
[15:47] me login or this like you know specific credential
[15:50] key and i don't really want to mess around with
[15:51] it i'm like look it's just logged in on my browser
[15:53] control my computer screen uh on my browser go
[15:56] click and figure it out yourself and it can go
[15:57] do that So that is one way to get around some
[15:59] of these MCPs and logins. But if you don't want
[16:02] to have to do that, and if you're using a lot
[16:03] of different things, or you have repetitive workflows,
[16:05] getting the MCPs and having it connect with them,
[16:08] number one uses less tokens for you because computer
[16:10] use where it's like looking at your computer
[16:12] and moving it around and screenshotting it all
[16:14] the time. That's super, it uses a ton of tokens.
[16:17] So if you want to use less tokens, I would get
[16:19] the MCP. If you want to make it more consistent,
[16:21] I'd get the MCP. It has less issues and glitches.
[16:24] And if you're doing, you know. pipelines and
[16:26] other things that are repeated tasks, I would
[16:28] definitely get MCPs for anything you're using.
[16:30] All right, I hope that was a useful episode.
[16:31] If you want to get access to the AI box MCP,
[16:34] it's $8 a month over on AI box .ai, you sign
[16:37] up for an account and then in your settings,
[16:39] you can grab the MCP link, you can go to Claude,
[16:42] give that to Claude and it will get that. Get
[16:44] that set up so you can access any of the tools
[16:46] you build on AI box and you get 80 plus different
[16:48] AI models now built into Claude or chat GPT or
[16:51] Gemini using that. So if you want to go check
[16:54] that out, it's AI box dot AI slash MCP. That's
[16:58] a page I created that will explain a little bit
[17:00] about how it works and how you connect it. There's
[17:02] a little diagram on there. Hope that is incredibly
[17:04] useful. Unlocks all the capabilities of what
[17:06] you'd like to do with Claude. Hope you guys all
[17:08] have a fantastic rest of your day and I'll catch
[17:10] you all in the next episode.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.