Windsurf Wave 4

For a brief period, we wondered whether there was so much in Wave 4 that we should push some of these capabilities to later waves. But that would be silly. Our users deserve the best we have to offer, as soon as we have it available.

The tl;dr highlights of Wave 4:

• Previews: Iterate on your apps rapidly by easily sending elements and console errors back to Cascade as context
• Tab-to-import: Another addition to our passive predictive experience
• Linter integration: Cascade checks its outputs against linters
• Suggested actions: After responding, Cascade suggests what your next action should be
• MCP discoverability: Simpler methods of identifying and pulling in helpful MCP servers
• Drag & drop files: Attach editor tabs or files from your file explorer to Cascade
• Model options admin control: For enterprises, specify which models your org can use
• Claude 3.7 Sonnet Improvements: A bit less aggression towards tool calling
• Referrals: Give the gift of Windsurf (and get rewards in return!)

You can read the full changelog here.

If you are new to the Windsurf family, we group a set of major feature releases together in a single update that we call a Wave. Wave 4 has a number of features unique to an AI-native IDE, marking a turning point from “catching up” to “leading the industry.”

There’s a lot here, so let’s unpack Wave 4…

Previews

It is very clear to us that many people are vibe-coding entire apps, but there are friction points. You have a local deployment of the app, but if something doesn’t work / look right, there are a number of manual steps you might have to take:

• You might try to explain what looks wrong or what is not working correctly in a prompt, hoping that Cascade will properly identify the component that you are referring to before making the edits.

• You might go into the browser’s console logs to copy any runtime errors, pasting them into Cascade’s input box to let it know what is going wrong.

Previews solve all of this. Instead of just showing you your application when you “locally deploy” your in-progress application, we add listeners and UX so that Cascade can be aware of exactly how you want to iterate:

• Point and click on the component that you want to make changes to
• Button to automatically pull console errors

This context is automatically passed back to your Windsurf Editor so that there is no manual copy-pasting or uploading necessary.

Previews should work with most web projects, React or otherwise, but not non-HTML websites such as those using WebGL or full-canvas screens. As of now, Previews are optimized for Chromium-based browsers and within the IDE, but should also support Safari and Firefox on most operating systems.

Tab to Import

While Cascade is the cornerstone of the Windsurf Editor, the passive predictive AI experience in the text editor is also critical to speeding developers when it is a matter of point edits. The evolution has been rapidly accelerating:

• Autocomplete (introduced 2022): predicts additions at your cursor position that could be accepted with the “Tab” key
• Supercomplete (September 2024): use the “Tab” key to not just accept additions at the cursor position, but entire nearby edits (additions/deletions/refactors) irrespective of cursor position
• Tab-to-jump (Wave 3): the “Tab” key could now also jump the cursor position when that was the appropriate next intent, as opposed to being restricted to code changes

In Wave 4, we are adding the “tab-to-import” functionality, where the “tab” key can be used to add imports to the top of files when a new dependency is used in the file:

Linter integration into Cascade

This has honestly been one of our biggest requests - if Cascade generates code that doesn’t pass a linter, then it should just fix itself first rather than be prompted to do so. Cascade now uses linter errors as an additional signal in its errors.

If Cascade generates code with linter errors and then takes a step to fix said generated lint errors, to the best of our ability, we will not charge flow action credits for the fix. This may be imperfect, but that’s not a reason to not try to do the right thing!

Suggested actions

Cascade is what we call a flow, or a collaborative agent. This means that it will return back to the conversation with you as the user to get your input on how to proceed, often when there are multiple reasonable next steps. Now, Cascade can suggest what these next steps might be so that you can stay even more tightly in the flow.

MCP Discoverability

In Wave 3, we launched our MCP integration, and it was generally well received, except for one minor nagging question - “what exactly is MCP and where can I find these MCP servers you speak of?” So we built a lot of educational content such as this, this, and this, but we wanted to bring a new level of discoverability for useful MCP servers directly in the product. We believe that the MCP integration is a key piece of the puzzle to unlock even more flow state with Cascade:

Drag & drop files

You asked, we delivered. Either drag in editor tabs or files from your file explorer to the Cascade input box, and we’ll attach the information to the prompt as context:

Admin Control for Model Options

As you may have noticed, we have added a lot more model support across many providers to Cascade over the last few weeks. The race on the foundation model side is intense, and we want to always provide the most optionality for our users so that they have the option for the highest quality results for their particular use cases. While we do have zero data retention with all of our model providers, we recognize that certain organizations may have strict guidelines on which models and providers have been approved for internal use. Now, administrators of Codeium Teams and Enterprise accounts can centrally set the availability of different models for the organization, another step in being the enterprise-ready AI platform:

Claude 3.7 Sonnet Improvements

Speaking of models, we were very excited to launch Claude 3.7 Sonnet and Claude 3.7 Sonnet (thinking) from Anthropic. That being said, we recognized that this model, as opposed primarily to Claude 3.5 Sonnet, was often a bit trigger happy with tool calling, leading to much-faster-than-anticipated usage of flow action credits. We’ve done a lot of work since the model release to reduce that tendency while still maximizing on the underlying strengths of this foundation model.

Referrals

Already, many of you are telling your friends and coworkers about Windsurf, and we sincerely thank you for inviting more people along this journey. That being said, we want to make sure we recognize you in some way that is more than just a thank you, which is why we are officially rolling out a referral program.

If you are a paying user, you can now go to https://windsurf.com/refer to find a personalized referral link. When someone who uses your link subscribes to a paid plan, both of you receive 500 free flex credits. There is no limit to how many people you can successfully get credit for referral, so refer away!

And as a reminder, some other things that have happened since Wave 3 that were not part of any Wave:

Unlimited Deepseek-v3

That’s right, we made DeepSeek-v3 cost zero user prompt and zero flow action credits on any paid plans. We still don’t believe that it is nearly as good as Claude 3.5 Sonnet and others for the very specific tool-calling task that we utilize these models for, and it is probably slightly worse than Cascade Base right now, but if we can make things free, we will!

Claude 3.7 Sonnet, Claude 3.7 Sonnet (thinking), GPT 4.5

We are big fans of model optionality for our users, as it is unlikely for there to be any model that is objectively better than all other models across every use case. So, as a list of powerful models have been released by the foundational labs in the last few weeks, we’ve made sure we’ve rapidly brought them to Windsurf. Again, we use these models for very specific tool-reasoning tasks, so a lot of the public benchmarks may not be reflective of how well these models may work for Cascade specifically, but this is also just a moment of time - any of these can get really good at any moment.

We have a lot more exciting updates right around the corner, so stay tuned!

And as a constant reminder, this is an enterprise ready solution, being used by software engineers from the smallest startups to the largest Fortune 500. We already support over a thousand enterprises in production, so ask your team to contact us to get access.

Surf’s up.