Many people either don’t care or know about the importance of context for LLMs and use the AI tools with vague/short queries like:
“Hey, build me the X tool” or “Generate a long-form article on Global Warming”.
This was me at some point and I got frustrated by the generic-slop responses I get from the models. However, the issue extends beyond the prompt, I needed a more systematic solution.
So I began developing the AI capabilities for my agency's management tool, Overbooked.
After hours of testing and iteration, I figured the output quality you can get with useful data plus agent tools is unmatched and I wanna share how OP the workflow gets.
There are two sides to the coin:
Using the MCP Server inside the IDE (Cursor for me)
You can connect Overbooked to modern IDEs like Cursor super easily with just this config:
{
"mcpServers": {
"overbooked": {
"url": "https://www.overbooked.app/api/mcp"
}
}
}Your Cursor agent now has access to over 15 valuable tools. Here are some real-life examples of how users, including myself, are utilizing them:
1. Juggling between related but separate repositories
I have many repositories that I actively maintain and Cursor has context only for the currently open project. This is fine but sometimes those projects are related (e.g. a NextJS frontend app coupled with a separate backend repo). When this is the case, you can either spend hours trying to put them into a monorepo or use the Overbooked MCP server like:
“Create a resource in project Alpha including the required changes we need to apply on the server repository.” and on the server repo:
“We need to do some changes on the X route, get the Required Changes resource and start refactoring.“
and just like that you preserved the context of both projects.
2. Delegating tasks and preventing hallucinations
You want to add a new feature to your app so you open Cursor and it starts implementing it. But it can get ahead of itself quite often, generating unnecessary code or losing track of the task. Instead, try this approach:
“We will implement X feature using existing modules whenever possible. Please create the necessary tasks and update their statuses as you go.”
The plan is now clear and accessible for the LLM. It also reduces the amount of hallucinations significantly.
3. Initializing and structuring new projects
Overbooked provides a variety of valuable tools for generating essential project documents, such as PRDs, tech stacks, and core features, as well as creating branded assets like color palettes, typography styles, and logos in seconds. You can always export them from the app, but they are also accessible via the MCP server for convenience. Try these:
“Initialize a Nextjs app based on the documents of project X.“
“Pull the styles of project X and update the global.css to include the colors and typography styles.”
Using the Chat Window inside the app (Overbooked)
All MCP tools are accessible within the project dashboard, and some of them are more effective when used in the app's chat window. Some examples:
1. Brainstorming in a group chat
Each project in Overbooked includes a dedicated group chat tab. This is very useful for brainstorming purposes because anyone on the team can tag Overbooked at any time to ask questions, gather information from the internet or plan actionable steps to move the project forward.
2. Generate long-form articles or social media posts
Overbooked can use project context to produce well-crafted long-form articles and social media posts. Below is a screenshot demonstrating Overbooked conducting research before generating the article:
3. Talk to your calendar
You can even talk to your calendar and let Overbooked create reminders for you.
If you wanna see it in action check out this demo we have prepared:
Overbooked Groupchat Agent Demo
Using this daily has boosted our productivity, lowered stress levels, and provided us with more free time. I invite you to explore it and share your thoughts; any feedback is welcome!