This needs to happen primarily not because of feature requests (file/var names can be difficult to pronounce/infer), but because of so much context that can be passed to models.

If we are saying "but the API for feat ABC is not 123 but XYZ", having an automcomplete that would finish it would be great.

Building on this, I think there are several improvements that could make the IDE more intuitive and powerful. For instance, automatic indentation and error correction should be standard when using Python files. If the IDE could learn from repetitive mistakes, it would significantly improve efficiency.

One example I often run into is with PyQt6 models, where the QAction import is frequently placed in the wrong location — instead of being in the GUI module, it’s mistakenly placed elsewhere. This is a recurring, avoidable mistake that not only wastes time but also disrupts workflow. The IDE should be able to recognize these frequent errors, flag them, and offer the correct fix so that developers never have to manually correct them again.

smart pip and enviroments compatibality ..ect

Additionally, auto-formatting and documentation generation should be built-in features. These aren’t just conveniences; they should be part of the development process. Templates for different kinds of files would also be a huge boost. If the IDE can enforce a consistent template for code, we’ll all end up with more organized, maintainable projects.

In short, the IDE should adapt to its users by automating these repetitive, time-consuming tasks, learning from common errors, and offering reusable templates that save time while ensuring best practices.--- windsuf coduim should employ me to uilt them the ide->>> i would superchatger it like no other and realy set the ide miles above the rest!!!! may the dev team need to contact me ...

Exactly, Sean — I’ve been thinking the same thing. If Windsurf already supports plugin integration for VS Code, then why doesn’t it expose the same kind of functionality internally? For example, we should be able to pass model inputs/outputs through an intermediary .py file, enabling seamless integration of text-to-speech (TTS), speech-to-text (STT), and other automation tools. This would massively boost coding flow and accessibility.

Why are we locked into MCP when we could be running lightweight custom Python scripts for things like local indexing or embedding? The IDE’s own indexing is extremely resource-intensive — it drags down performance, and there’s no option to offload or run this more efficiently. Let us use our own embedders and scripts. Local embedding is not only possible, it’s already common in open-source workflows.

There’s also no obvious way to save or export chats — at least not on macOS. That’s a fundamental feature, and it feels intentionally restricted. Why lock us out of our own data?

All of this makes Windsurf feel like a business-focused, locked-down product, not a developer-centric one. Developers should be able to bring in their own models, use local resources, and extend the platform with custom tools — just like we can in VS Code with plugins like DeepSeek, OpenDevin, or even Quinn. Many of these are free and already support model bridging or CLI chaining.

Windsurf has so much potential, but these restrictions are limiting. Give us an API, plugin SDK, or even just a configurable execution bridge. We don’t need to depend entirely on MCP if we can run our own code more efficiently. And we want to — because we care about stability, control, and performance.

There is a voice to prompt feature in the Windsurf IDE for the Cascade input. I had looked into building this exact same thing because the TTS and STT tools are already available elsewhere and can be quickly integrated. I started building a custom VS Code extension to try and hook into Cascade and essentially bridge the Cascade commands to terminal command line. If the tool had CLI, this would allow for chaining of other tools with Windsurf, including TTS and STT from a mobile context so we can develop on the go (coding while biking or running would be great)