What problem are you trying to solve?

Focus on the what and why of the need you have, not the how you'd like it solved.

While Compass currently features a helpful natural language query builder, it is limited to translating text into filter objects or aggregation pipelines that the user must then apply. It does not execute operations autonomously. Developers often need to perform multi-step database interactions, CRUD operations, or administrative tasks that still require manually writing complex mongosh commands. Bridging the gap between a user's natural language intent and the actual execution of those commands remains a manual bottleneck in the GUI.

What would you like to see happen?

Describe the desired outcome or enhancement.

I propose adding an "Agentic AI" tab or extending the embedded shell with an "Agent Mode." Instead of just generating a query string, this feature would utilize an LLM with tool-calling capabilities to autonomously determine the necessary database actions. The agent would construct the exact Node.js driver/mongosh commands, execute them safely in the background, and return the formatted results directly to the user. I have successfully prototyped this architecture using PyMongo and tool-calling, and I am interested in contributing a TypeScript/Node.js implementation of this feature to the open-source Compass repository.

Why is this important to you or your team?

Explain how the request adds value or solves a business need.

An autonomous agent dramatically accelerates database workflows. It allows developers to interact with their databases seamlessly without constantly context-switching to look up precise mongosh syntax. Evolving Compass from a query translator into a highly advanced, automated database assistant will significantly boost productivity and make complex database management much more accessible.

What steps, if any, are you taking today to manage this problem?

Currently, I rely on a mix of manual workarounds. I built a custom external Python script using PyMongo and an LLM to autonomously generate and execute database operations from my natural language prompts. However, when working within Compass, I have to constantly context-switch between the GUI, external AI chat tools to generate my queries, and the embedded mongosh terminal to manually paste and execute them. This disjointed workflow forces me outside of the Compass environment and slows down routine database management.