Qoder in Plain English: A 3,000-Word Guide to Your New AI Coding Partner

❝

Cover image:

❞

Table of Contents

1. What Exactly Is Qoder?
2. Why “Agentic” Beats Ordinary Autocomplete
3. Five-Minute Setup: From Download to First Commit
4. Three Core Features You’ll Use Daily

   • Next-Edit Suggestions (NES)
   • Inline Chat
   • AI Chat (Ask vs. Agent)
5. Going Further: Memory, Rules, and MCP
6. How Credits and Pricing Work (Today and Tomorrow)
7. Quick-Reference FAQ
8. Next Steps & Best Practices

1. What Exactly Is Qoder?

「Pronunciation first:」 say “KO-der.”

Qoder is a desktop coding environment that wraps an AI agent around your entire codebase. Instead of guessing the next line, it studies the whole project—files, folders, dependencies, even your personal style—and then offers edits, explanations, or full feature implementations.

Think of it as three tools in one:

• A 「code-aware editor」 (like VS Code)
• An 「AI pair programmer」 that remembers context across sessions
• A 「task runner」 that can write, test, and commit code end-to-end

❝

Key takeaway: Qoder does not just autocomplete; it collaborates.

❞

2. Why “Agentic” Beats Ordinary Autocomplete

「Analogy:」

• Traditional tools are like a spell-checker.
• Qoder is like a senior teammate who skims the whole document, rewrites awkward paragraphs, and then runs the tests to make sure nothing broke.

3. Five-Minute Setup: From Download to First Commit

3.1 Install

1. Visit https://qoder.com/download and grab the installer for macOS, Windows, or Linux.
2. Double-click the file and follow the on-screen wizard.
3. Launch Qoder; you will see a dark/light theme toggle on first run.

3.2 Sign-In

• Click the user icon (top-right) or press

  • macOS: ⌘ ⇧ ,
  • Windows: Ctrl Shift ,
• Choose 「Sign in」.
• Use Google, GitHub, or email registration.
• Return to the IDE; the status dot turns green when you are authenticated.

❝

During Preview, all features are free. Paid tiers will appear “in the coming weeks,” according to the official changelog.

❞

3.3 Open or Clone a Project

「Option A: Open Local Folder」

• macOS: ⌘ O
• Windows: Ctrl O
• Select any folder with code.

「Option B: Clone from GitHub」

1. Click 「Clone repo」.
2. Paste a URL or pick from the GitHub browser.
3. Choose a local parent directory and hit 「Clone」.

4. Three Core Features You’ll Use Daily

4.1 Next-Edit Suggestions (NES)

「Purpose:」 AI shows what to change, add, or delete 「right where your cursor is」.

How to Turn It On

1. User icon → 「Qoder Settings」 → 「Completion」 → Toggle 「NES」.
2. In any file, type a partial line or a plain-English comment, e.g.

   # create a dataclass for user profiles
   

3. Press ⌥ P (mac) or Alt P (win).
4. Ghost text appears in italics.

   • Tab accepts the full suggestion.
   • ⌘ → (mac) or Ctrl → (win) accepts word-by-word.
   • Esc discards.
   • ⌥ [ or ] cycles through alternatives.

Supported Scenarios

❝

When the edit lands outside your viewport, 「Tab to Jump」 appears. Press Tab to teleport to the exact line—no manual scrolling.

❞

4.2 Inline Chat

「Purpose:」 Highlight a block of code, ask a question, or give an instruction; receive an immediate rewrite inside the file.

Steps

1. Highlight any code (or place the cursor on a line).
2. Press ⌘ I (mac) or Ctrl I (win).
3. A small chat box opens. Examples:

   • “Convert this loop to a list comprehension.”
   • “Add type hints and a docstring.”
4. Hit ⌘ ⏎ (mac) or Ctrl Enter (win) to apply the diff.

Typical Uses

• Refactor repetitive snippets
• Generate unit tests for the selected function
• Explain cryptic legacy code in plain English

4.3 AI Chat (Ask vs. Agent)

Open the right sidebar with ⌘ L (mac) or Ctrl L (win). You will see two modes at the top.

Ask Mode Example

Prompt:

How do I read a JSON file safely in Python?

Response: a concise code snippet plus explanation. No files change.

Agent Mode Example

Prompt:

Add a REST endpoint /ping that returns {"status":"ok"} in FastAPI.
Write tests and run them.

Agent’s workflow:

1. Searches the codebase for an existing FastAPI app.
2. Creates or modifies main.py.
3. Creates test_main.py.
4. Runs pytest in the built-in terminal.
5. Reports success or failure in chat.

Reviewing Changes

• Each modified file shows 「Generating → Applying → Applied」 states.
• Click 「View Changes」 to open a diff viewer.
• Accept or reject per file or per hunk.
• Use 「Undo」 to roll back at any point.

5. Going Further: Memory, Rules, and MCP

5.1 Long-Term Memory

Qoder stores three layers:

1. 「Personal style」: indentation, quote style, favorite libraries.
2. 「Project rules」: folder conventions, lint rules, README templates.
3. 「Historical context」: earlier chats, accepted edits, rejected patterns.

Manage it under 「Qoder Settings → Memory」. You can add plain-English rules such as

❝

“All public methods must include a docstring with example usage.”

❞

The agent will follow them in every future session.

5.2 MCP (Model Context Protocol)

MCP lets you plug external tools into Qoder—think databases, internal APIs, or custom linters.

• Install from the built-in marketplace with one click.
• Configure in 「Qoder Settings → MCP」.
• The agent can then query your staging database or run bespoke security scans without additional setup.

6. How Credits and Pricing Work (Today and Tomorrow)

❝

Source: qoder.com/pricing

❞

• Credits measure usage of Chat and Agent requests.
• Credit Packs will be sold à-la-carte if you run out mid-month.
• When credits hit zero, the agent falls back to a basic model instead of stopping work.

❝

During Preview, everyone is on the 「Free」 plan plus a 14-day Pro trial. No payment method is required today.

❞

7. Quick-Reference FAQ

「Q1: Which programming languages are supported?」
The official docs do not list an exhaustive set, but everyday tests show solid support for Python, TypeScript, Go, and Java.

「Q2: Can I use it behind a company firewall?」
Initial sign-in and cloud model calls need internet. Afterwards, caching reduces external traffic. Full offline mode is not yet available.

「Q3: Will Qoder replace my current IDE?」
It is a full standalone editor, yet keyboard shortcuts and layout will feel familiar to VS Code users. Settings sync and extensions are on the roadmap.

「Q4: What happens after the free preview ends?」
You may:

• Upgrade to a paid tier; or
• Stay on the free tier with limited credits.

「Q5: How is this different from GitHub Copilot?」
Copilot focuses on inline suggestions; Qoder adds autonomous multi-file edits, command execution, and persistent memory.

「Q6: Can I write my own MCP plugins?」
Yes. The protocol is open, and documentation will be released alongside the public SDK.

8. Next Steps & Best Practices

1. 「Try NES on your oldest file.」
   Open legacy code, press ⌥ P / Alt P, and watch the AI propose modernizations.

2. 「Teach the agent your style guide.」
   Add one rule per day in 「Settings → Memory」; after a week you will see consistent formatting without asking.

3. 「Clone a sample repo.」
   The official quick-start guide recommends cloning a small open-source project to test Agent mode end-to-end.

4. 「Monitor credit usage.」
   Check the bottom-left counter so you are never surprised when pricing launches.

5. 「Follow the changelog.」
   Bookmark qoder.com/changelog for weekly updates on new MCP servers and editor features.

Final Word

You now have a complete, hype-free map of Qoder. Install it, open a real project, and let the agent do the tedious parts while you focus on the big picture. Happy coding!