AI Coding Agents Might Not Need Tools Anymore

追加: 2026-05-27

[00:00:00]Imagine you're a 19th century pirate and during your looting you've stumbled upon a new AI tool. Previously, you have been using MCP tools with your AI agents.

[00:00:08]When you have an MCP tool connected, your model has to call that tool and it gets some results back. The pirate has the Gmail MCP connected because he needs to email his therapist about his work-life balance. With the MCP, you can ask Claude to find emails from a specific person and reply to them. It calls the search tool to find them, then gets the threads to read each one, then the reply tool to send a response. But what if the Gmail MCP was literally a folder and Claude could go inside it the same way it goes inside any folder on your computer? Your emails would be actual files that Claude reads the same way it reads markdown files in your code base. That's exactly what Mirage is.

[00:00:42]It's basically a virtual file system for AI agents. But before we dive deeper into this virtual file system, let's take a look at why file systems are so important for these AI tools. The significance of file systems comes down to how AI models actually learn things.

[00:00:57]When you give Claude a custom tool, whether it's an MCP or an API, you're asking it to learn something new on the spot like a baby being forced to start walking at gunpoint. The agent has to read the description and plan which calls to make in what order. That's tokens of overhead spent before it does any actual work and it repeats for every tool you add. A file system doesn't have that problem because it's the one interface every LLM has already learned thoroughly during repetitive training.

[00:01:22]Basic file commands have been used billions of times in the code the model trained on. Unix has been refining this for 50 years and it's why file systems are the one abstraction that lets an AI agent work across multiple services without repeatedly forgetting its life's purpose. During thorough testing, I found lots of sneaky little things that aren't written inside the tool's GitHub description. After the setup, we asked Claude to mount Gmail to the Mirage workspace. Like a good little boy, it had to guide me through the Google Cloud Platform for the OAuth creation so that we could get the credentials. When you start Claude inside the Mirage folder, you'll be able to see the services you mount. You can see that Gmail has been added here. And for each Gmail label and category, it could see a different directory. Now, before we dive into the new features, we got to put in the word from our sponsor. Before you point the finger of judgement, remember the editors' kids need to eat, too. Now, if you're paying for ChatGPT, Claude, Gemini, and Midjourney separately, you're already spending way more than you need to. Chat LLM from Abacus AI brings over 100 AI models into one platform, including ChatGPT 5.5, Claude Opus 4.7, Gemini 3.1, Grok, and Deep Seek. And every new model gets added instantly. You don't even have to choose because Root LLM automatically picks the best one for your prompt. But it goes beyond just chat. You can use it for writing and deep research, generating images and videos from top models like C-Dance 2.0 and Nano Banana. They also have the Abacus AI agent. By chatting with the AI, you can build professional presentations, create full-stack apps and websites that even accept payments, so you can launch a business without writing a single line of code. So, chat, deep research, images, videos, apps, websites, AI agents, presentations, literally everything in one platform.

[00:03:02]All of this for just $10 a month instead of paying for separate AI subscriptions.

[00:03:06]Check it out at chat.abacus.ai or click the link in the description and start building now. Now, there's a problem with the existing Gmail MCP. Emails regularly get file attachments with them, but using the MCP, Claude can't read or download them, which pretty much makes it a blind old man in this situation. It can only see the file name, which creates a serious limitation with the automation. But with Mirage, if we ask it to search for any attachment, what's going to happen is that it can actually read the file attached to the email because it's just a file in the file system. With the MCP, you have to get the attachment manually and paste it into the agent's context manually. In our inbox, it found this extremely legal invoice, and it read the whole thing and gave us its actual content. Unlike your ex, Mirage actually cares about you and gives you some predefined tools that you can easily add such as notion, the whole Google suite, Telegram, Slack, and even different storage systems or databases as well. If you connect them, they will appear as folders to Claude in its directory. And the same thing can happen with Slack where chats become files and any file attachments that your team sends there can also be accessed easily. Since Google Drive is a storage platform, its MCP will obviously give you a download tool for the files, but there's a huge flaw in the download mechanics. thing is only efficient for small things. And no, we aren't making that joke here because my grandmother watches these videos. For example, we had this small 2 MB file that we wanted the Google Drive MCP to download and place somewhere else. To download the file, it fetches the file's content as a string and then it is decoded and written back into your file system. Even this request took about 4 minutes and also bloated the context.

[00:04:42]And if the file size is huge, let's say 100 MB, then it would hit the response limit and waste the context as well.

[00:04:49]Mirage solves this because you can just copy the files from the drive onto your system. For example, we had this sample test data file. It could actually read it and it used the copy bash command to just copy it into my documents folder.

[00:05:01]The setup [snorts] process for this is actually pretty long, so we're not going to go through each step and we don't need to because we're not living in the olden times anymore where you actually had to read documentation like a common peasant. What we recommend is just cloning the source code directly onto the system. And once you've cloned the source code, just open Claude inside that and let it guide you through the installation. Going through the installation steps one by one doesn't really matter because it's all code and Claude can automatically do everything.

[00:05:27]But if you run this on a Mac, the install isn't that straightforward because God forbid Apple lets you install anything without feeling like you're about to burn your system to the ground. To use Mirage, macOS needs to have it mounted as a file system so that your agent can use it as just another folder with files. This is why it uses this library called MacFUSE. Using this software, macOS starts supporting third-party file systems. Claude can automatically download it via Homebrew, and all you have to do is follow its steps. Basically, MacFuse is a system extension, and you do need to restart your Mac and change the security configurations. So, that is something that you should be aware of. But, for those of you who are living under a rock and are new to AI agents, this might cause a problem for you. When it's going to restart, that chat session is going to be lost as well. So, either rename that chat session using the built-in / command or ask it to just read the previous chats because all of them are stored on the system. This is why when we restarted the system, we had to give it the chat again, and it recalled the context that the chat was about setting up Mirage for the video. Most of the tools that you will use need to be authenticated. Unlike the connectors for Claude, automatic OAuth isn't built in here. So, whatever tool you need to install, first it's going to mount a TypeScript file where it's going to add the tool as a directory in this workspace. After that, you're required to add in the credentials that the tool needs. For Google applications, you will have to go into the Google Cloud Console, which has the user experience of filing taxes in a foreign language.

[00:06:52]You have to enable APIs and get credentials, and that might take you some time. It is a little time-consuming, but Claude is going to guide you through it pretty easily.

[00:07:00]After all of this is done, Claude is going to ask you to run a mount command in a separate terminal and keep that terminal open. Now, if you don't want to keep the terminal open, you can ask Claude to run it as a background process. And after that, you basically just go into the Mirage folder, start Claude there. And as you can see, when we list the files, it has all my connectors, and it can actually look inside them and search for files. Now, if you've been following the channel for some time, first of all, thank you.

[00:07:24]You're our favorite human being. And you probably remember the Code Mode video on Docker, where Docker's Code Mode could basically make custom tools by combining tools from different MCPs so that repeated operations could be run easily.

[00:07:36]This is the same thing. You can do the same just by using bash commands combined together so that the contents of one file can be copied into another.

[00:07:43]For example, here what we asked it to do was find emails in the inbox that mention sponsors, get us the top three sponsors, and yes, top three obviously means the sponsors that give us the most dough. Then it would list them on the Notion page. And another thing that code mode or TypeScript tools in general solved was that context never really had to enter the model's context window. For example, here the model just has to write bash commands that fetch one thing from one file and then put it into another. The model doesn't need to have all of this loaded context placed inside its context window. So, the same thing happens here with bash commands. They form a pipeline, and they can easily connect these different tools together.

[00:08:20]Now, these [snorts] are not the only tools that you can use with this virtual file system. What's basically happening on the back end is that it's using APIs from these tools, the same APIs that are used by the MCP servers, and it's just turning those into file systems. This is like convincing the homeless man outside our building that he is the chosen one, and he can actually fly. So, the same can be done with any custom service that you want. If you want a Figma MCP server where all the Figma designs are listed as files, you can do that as well.

[00:08:46]Although, we don't think people honestly use Figma that much anymore. If we recall correctly, it was supposedly just killed by Claude design weeks ago. But other than that, we tried it out on the Google Chat APIs. Internally in our team, we use Google Chat, and yes, we want you to judge us for that. Google Chat allows you to go into spaces and add different apps there. So, we basically made this test space, and inside it we added this Mirage app. It basically gave it access to read everything in there. And as you can see, it automatically fetched everything and was able to read all the context, and you can do this easily with any service.

[00:09:18]Again, as we mentioned in the setup, make sure that you have the repo installed locally, and that Claude can read everything in there, so that it doesn't have any gaps in its context. Of course, it's going to take some retries because just like us, Claude isn't perfect. You're going to get a lot of errors, and you might have to do multiple feedback loops with the agent, but you can pretty much build any tool into a file system. Now, [snorts] we already told you that you don't have to keep it running in a terminal. It can run as a background process, but every time that process restarts, you're starting from scratch. The cache and the index are what make the workspace persistent, and both of them are gone on every restart. This is exactly what every new AI founder has been doing for the past 2 years. Mirage fixes its own problem with a demon that turns it into a persistent background server. You can have multiple workspaces running at the same time. They can be named, and each one starts in whatever mode you want.

[00:10:07]But let's say you don't want this on your local machine, or you work across multiple machines because your fat home-schooled brother can't get enough of Roblox on the PC. But because it's a standard HTTP server, you can host it anywhere and switch to your M3 Max MacBook. That's what lets you turn Mirage into a file system knowledge base. That brings us to the end of this video. If you'd like to support the channel and help us keep making videos like this, you can do so by using the Super Thanks button below. As always, thank you for watching, and I'll see you in the next one.