Gallery

Contacts

405 W. Greenlawn Ave Lansing, Michigan 48910

contact@techjacksolutions.com

+1-616-320-4064

WINDSURF / DEVIN DESKTOP

How to Use Windsurf (Devin Desktop) & Cascade (2026)

How to use Windsurf (Devin Desktop): install the editor, open Cascade, and pick a model

To use Windsurf, you install it as a standalone desktop editor, sign in with a free account, and open its built-in agent, Cascade, with a single keyboard shortcut. That is the short version of how to use Windsurf, and the rest comes down to a few choices: whether to import your old editor setup, when to work in Code mode versus Chat mode, and which model to send each request to. Windsurf is an AI code editor, a full desktop IDE with autocomplete, syntax highlighting, and debugging, plus Cascade, the agent that can plan and carry out changes across your whole project. This guide walks the path end to end: installing the editor, onboarding, opening Cascade, working in its two modes, using tool calling and checkpoints, picking a model, and keeping an eye on your usage quota.

First, the name: Windsurf is now Devin Desktop, the editor built with Cognition. The change arrives as an over-the-air update, and your existing plan, pricing, extensions, settings, and in-progress work carry over automatically. Only the name and branding changed, so you may still see "Windsurf" in some menus, and the JetBrains plugin is still called Windsurf for JetBrains. Throughout this guide we use both names interchangeably.


Before You Start

Learning how to use Windsurf is quick. You download one application, sign in with a free account, and you are working. The thing to understand up front is that it is not an extension you add to an editor you already use. It is its own full IDE, so the first step is installing the app, not configuring a plugin. If you already run the editor, the Devin Desktop rebrand reaches you as an automatic update with nothing to reinstall.

Setup Checklist
A supported operating system – The desktop app runs on Mac (OS X Yosemite or later), Windows (10 or later), and Linux (glibc 2.28 or later). Download it from windsurf.com.
A free account – Sign in to start on the free plan. No credit card is required, and it includes unlimited inline edits and Tab completions plus a light agent quota.
Your current editor setup, if you want it – On first launch you can import settings, extensions, and keybindings from VS Code or Cursor, or start fresh.
The right plan for your goal – The free plan is fine for learning. Heavier agent use, the full model lineup, and free SWE-1.6 sit on Pro and above.

One thing to settle early: Cascade can read your whole project and run tools on your behalf. Before you point it at a private repository, add a .codeiumignore file for anything it should not touch. We cover that in the usage step.


How to Use Windsurf: Install and Onboard

Download the desktop app for your operating system from windsurf.com and install it like any other application. There are builds for Mac, Windows, and Linux, so the same editor and the same Cascade agent are available wherever you work. Because the editor is a fork of the familiar VS Code base, the layout and shortcuts will feel immediately recognizable.

The first launch is the moment that saves you the most time. Windsurf offers to import your settings, installed extensions, and keybindings directly from VS Code or Cursor, so your environment arrives configured rather than blank. If you would rather have a clean slate, you can skip the import and start fresh. Either way, a free account is enough to begin, and you can sign in and be in the editor within a couple of minutes.

Import
On first run, pull your VS Code or Cursor settings, extensions, and keybindings into Windsurf so your environment is ready immediately, or choose to start fresh.

Once you are signed in and a project folder is open, the AI features described below are all a keystroke away. The next step is opening the part that makes Windsurf an agentic IDE rather than a plain editor: Cascade.


Open Cascade

Cascade is the built-in agent, and it is the centre of the Windsurf workflow, so most of how to use Windsurf well runs through this panel. You open it with a single shortcut: Cmd+L on Mac, or Ctrl+L on Windows and Linux. That brings up the Cascade panel alongside your code, where you type instructions in plain language and watch the agent work.

What sets Cascade apart from a chat window bolted onto an editor is that it has real-time awareness of what you are doing in the editor, and a feature called Fast Context that pulls in the relevant parts of your codebase so it reasons about the whole project rather than a single open file. Behind a request, Cascade runs a planning step and keeps a Todo list, breaking a larger instruction into the steps it intends to take before it starts editing.

How to think about it: Cascade is where you describe an outcome and let the agent work out the edits. Tab completions and inline edits still handle the small, mechanical typing. Reach for Cascade when the change spans more than the line in front of you.


Code vs Chat Mode

Cascade runs in two modes, and knowing which to use is the single most useful habit for a newcomer. Code mode lets Cascade edit files directly: you describe a change and it makes the edits, which you then review and accept or refine. Chat mode is for questions and suggestions, where Cascade explains, plans, or proposes without touching your code.

When to use each

Use Chat mode when you are exploring, learning a codebase, or you want a plan before any change is made. Use Code mode when you have decided what you want and you would rather Cascade carry out the edits than type them yourself. A common rhythm is to start in Chat to agree on an approach, then switch to Code to have Cascade implement it.

Step What You Do
1. InstallDownload Windsurf (Devin Desktop) from windsurf.com and sign in with a free account
2. OnboardImport your VS Code or Cursor setup, or start fresh, then open a project folder
3. Open CascadePress Cmd+L or Ctrl+L to bring up the agent panel
4. Choose a modeUse Chat to plan and ask, Code to let Cascade edit files
5. Pick a modelUse SWE-1.6 to conserve quota, or a frontier model for harder work
Guide Progress
0 of 8 sections complete

Tool Calling and Checkpoints

In Code mode, Cascade does more than write text into files. It calls tools to get its work done: it can search and analyze your code, run a web search, execute commands in the terminal, and reach external systems through MCP servers. A single prompt can trigger up to twenty tool calls, and with Auto-Continue enabled Cascade keeps going through a multi-step task without stopping to ask after each step.

That autonomy is what makes Cascade fast, which is exactly why the second feature matters. As Cascade works, it creates checkpoints, named save points you can return to. If a change goes in a direction you did not intend, you revert to an earlier checkpoint and the project returns to that state, rather than untangling the edits by hand. The practical workflow is to let Cascade run, review what it produced, and revert cleanly if you want to try a different approach.

Up to 20
Cascade can make up to twenty tool calls in a single prompt, and Auto-Continue lets it carry a multi-step task through without pausing. Checkpoints give you a clean way back if it goes off course.

For a first session, leave a task small and review the checkpoints as you go. Once you trust how Cascade behaves, longer Auto-Continue runs and bigger tasks follow naturally, with checkpoints as your safety net. If you want the bigger picture before going deeper, our breakdown of what Windsurf is covers the editor and the Cognition rebrand in full.


Pick a Model

Windsurf is not tied to one AI provider, and the model you choose is partly about capability and partly about how fast it burns your usage. The default fast option is SWE-1.6, Cognition's own coding model, which is free on the Pro plan and above. It is built for quick, agent-driven coding, so for most everyday work it both performs well and conserves quota.

For harder problems, the paid tiers also give you frontier models from OpenAI, Anthropic (Claude), and Google (Gemini), which suit deeper reasoning and trickier changes. At the other end, there are smaller and open models, such as Claude Haiku, GPT-5.2 Mini, and Kimi K2.5, that you can switch to when you want to stretch your quota further on lighter tasks. SWE-1.6 is Cognition's model, while the frontier and open models are third-party, so attribute them accordingly.

On model choice: match the model to the job. Reach for SWE-1.6 or a mini model for routine edits to save quota, and step up to a frontier model only when a task genuinely needs deeper reasoning. The available list changes, so confirm current options in the in-app model picker.


Manage Usage and Quotas

Each plan includes an amount of agent usage that refreshes on a daily or weekly basis, and the cost of a message varies by the model you pick and the work involved. Because of that, the single most useful habit is to be deliberate about which model you send a request to: heavy use of a frontier model will draw down quota far faster than the same work on SWE-1.6 or a mini model.

The free plan is generous for learning, with unlimited inline edits and Tab completions and a light agent quota. As of June 16, 2026 the vendor lists the paid tiers as Pro at $20 per month, Max at $200 per month, and Teams at $80 per month plus $40 per developer seat, with Enterprise available through sales for VPC, SSO, and dedicated support. On Pro and above you can buy extra usage at API rates if you run past the included amount. These figures are vendor-reported, so confirm current pricing before you buy.

Control what Cascade sees

To keep secrets, generated output, and other sensitive paths out of the agent's reach, add a .codeiumignore file to your project. It works like a familiar ignore file: anything you list is excluded from what Cascade reads and acts on, which is worth setting up before you point the agent at a private repository.

That is the whole of how to use Windsurf for a productive first week: install it, import or start fresh, open Cascade, switch between Code and Chat as the task demands, pick a model that fits the work and your quota, and lean on checkpoints. The best next step is simply to install the free plan and run a small task end to end. To weigh it against the other leading standalone editor first, see the comparisons in the Related Reading below.

Cost per message varies, so watch the model
Usage is metered as a quota that refreshes daily or weekly, and the cost of each message depends on the model and task. There is no single per-message price to memorize, so treat model choice as the main lever and check your usage in the app.
Set .codeiumignore before the first real session
Cascade reads and acts across your project, so add a .codeiumignore for secrets and sensitive paths before you open a private repository, the same way you would manage any control over where source code is allowed to go.

Troubleshooting

These are the questions newcomers run into most often, with the practical answer for each.

Yes. Windsurf has rebranded to Devin Desktop, built with Cognition. The change arrives as an over-the-air update, and your plan, pricing, extensions, settings, and in-progress work carry over automatically. Only the name and branding changed, so you may still see "Windsurf" in some menus, and the JetBrains plugin is still called Windsurf for JetBrains.
The free plan includes unlimited inline edits and Tab completions plus a light agent quota and a limited set of models. If you hit the agent ceiling regularly, the Pro plan at $20 per month (vendor-listed June 16, 2026) raises the quota, makes SWE-1.6 free, and unlocks the full model lineup. You can also conserve quota by sending routine work to SWE-1.6 or a mini model rather than a frontier model. Confirm current limits at windsurf.com/pricing.
Use checkpoints. As Cascade works in Code mode it creates named save points, so when a change goes in a direction you did not intend, revert to an earlier checkpoint and the project returns to that state. This is cleaner than undoing edits by hand, and it is why reviewing as you go, and reverting freely, is the right habit in a first session.
Start with SWE-1.6, Cognition's fast coding model that is free on Pro and above, and change it per task. Step up to a frontier model from OpenAI, Anthropic, or Google for deeper reasoning, or down to a mini model such as Claude Haiku, GPT-5.2 Mini, or Kimi K2.5 to conserve quota on lighter work. Cost per message varies by model, so the picker is the source of truth for current options.
Fact-checked against vendor documentation and official sources, June 2026. Verify current pricing at windsurf.com/pricing before purchasing.
Windsurf and Devin are trademarks of Cognition AI, Inc. This article is an independent editorial resource by Tech Jacks Solutions. Not affiliated with or endorsed by Cognition AI, Inc.
Before You Use AI
Your Privacy
Windsurf, now Devin Desktop, processes the code, prompts, and context you give it to power Cascade and its other AI features, and frontier models route requests to their third-party providers. Add a .codeiumignore file to keep secrets and sensitive paths out of the agent's reach, and review the vendor's privacy and enterprise documentation, including SSO and VPC options on Teams and Enterprise, before working in proprietary repositories.
Mental Health & AI Dependency
AI coding agents like Cascade make it fast to accept changes you have not fully read, and with Auto-Continue running multi-step tasks it is easy to delegate more judgment than you intend. Treat generated code as a draft to review and test, use checkpoints to step back, and notice if the tool is replacing rather than supporting your own understanding. If you are experiencing distress:
  • 988 Suicide & Crisis Lifeline: Call or text 988
  • SAMHSA Helpline: 1-800-662-4357
  • Crisis Text Line: Text HOME to 741741
AI systems can produce plausible-sounding but incorrect guidance. For mental health, medical, legal, or financial decisions, always consult a qualified professional.
Your Rights & Our Transparency
Under GDPR and CCPA, you have the right to access, correct, and delete your personal data held by AI service providers. Tech Jacks Solutions maintains editorial independence from all vendors covered on this site. Some links may be affiliate links, which help fund independent research at no extra cost to you. The EU AI Act classifies AI systems according to their intended use and risk level.