Tool Guide · Coding
Continue Guide
Open-source AI coding assistant. Lives as an extension inside VS Code, JetBrains, or Cursor. BYO-model and full configuration control.
Key takeaways
- 1Works in VS Code, JetBrains, and Cursor
- 2Model-agnostic — any provider, any local model
- 3Tab autocomplete + chat + edit
- 4Configurable via JSON — granular control over models, prompts, slash commands
- 5Custom slash commands for repeatable tasks
- 6Indexes your codebase locally
- 7MIT-licensed, fully open source
Best for
- Engineers who don't want to switch editors
- Teams needing fine-grained control over models per task
- Privacy-sensitive workflows (local model + local index)
- JetBrains users (most polished AI extension for IntelliJ family)
Not for
- Engineers who want a single-vendor solution
- Workflows that demand strong autonomous agents (Cascade / Claude Code beat it here)
What it is
Continue is an open-source AI coding extension that installs into VS Code, JetBrains IDEs, or Cursor. It gives you autocomplete, chat, and inline edits — but unlike Cursor / Windsurf, you don't change editors.
The trade: Continue is more flexible (any model, any IDE, full config) but requires more setup.
Strongest fit
If you live in JetBrains (IntelliJ, PyCharm, WebStorm), Continue is the clear AI extension to install. The JetBrains AI Assistant exists but Continue's flexibility usually wins.
If you live in VS Code and don't want to switch to Cursor / Windsurf, Continue gives you 90% of the AI experience without leaving the official editor.
Configuration overview
Everything is driven by a config.json file:
{
"models": [
{ "title": "Claude Sonnet", "provider": "anthropic", "model": "claude-sonnet-4-7" },
{ "title": "Local DeepSeek", "provider": "ollama", "model": "deepseek-coder-v2:16b" }
],
"tabAutocompleteModel": { ... },
"slashCommands": [...],
"contextProviders": [...]
}
This JSON is the source of truth. It's more work than dropdown configs but gives you fine-grained control.
Custom slash commands
The killer feature for power users:
{
"name": "test",
"description": "Generate edge-case-aware tests",
"prompt": "Write unit tests for the highlighted code. Cover happy path, boundaries, invalid input, domain edge cases. Use {{test_framework}}."
}
Now /test does exactly what you want, every time.
Local-first option
Continue is one of the best paths to fully-local AI coding:
- Run Ollama with DeepSeek-Coder or Qwen-Coder.
- Configure Continue to use it.
- Index your codebase locally.
Quality is below frontier APIs but completely private.
When to skip Continue
- You want a polished out-of-the-box experience: Cursor / Windsurf.
- You want autonomous agent mode: Claude Code / Cursor agent / Cascade.
- You don't want to touch JSON: any of the GUI alternatives.
Continue is for engineers who value flexibility over polish, and want to keep their existing editor. For that audience, it's the strongest option.
Pros and cons
Pros
- Open source, model-agnostic, JSON-configurable
- Works inside your existing IDE — no switching
- Great JetBrains integration
- Local-first option for privacy
Cons
- Setup is more involved than Cursor / Windsurf
- Autonomous agent capabilities lag behind dedicated tools
- JSON config can feel intimidating for non-power-users
Prompt examples
Copy any of these, replace the placeholders, run.
Custom slash command (config.json)
{
"slashCommands": [
{
"name": "review",
"description": "Senior code review of the current file",
"prompt": "Act as a senior engineer. Review this file for: correctness, security, performance, maintainability. Output structured findings with line refs."
}
]
}Local Ollama model
{
"models": [
{
"title": "DeepSeek Local",
"provider": "ollama",
"model": "deepseek-coder-v2:16b"
}
]
}Alternatives
Frequently asked questions
Related on AIKnowHub
Tool Guide
Aider Guide
Open-source terminal coding agent. Brings AI pair-programming into git-tracked repos with surgical edits and built-in commits.
Tool Guide
Claude Code Guide
The definitive guide to Anthropic's terminal-native AI coding agent — install, configure, MCP, hooks, skills, sub-agents, plan mode, cost, security, limitations.
Tool Guide
Cursor AI Guide
The most popular AI-native IDE. A VS Code fork with deep model integration, multi-file edits, and a strong agent loop.
Comparison
Best AI Model for Coding
Which model actually writes the best code? Benchmarked against real tasks — refactors, bug fixes, new features, code review.
Directory
Aider
Open-source terminal pair-programmer. Per-commit AI history, model-agnostic, MIT license.
Directory
Claude Code
Terminal-native agent that reads, writes, runs, and reviews your codebase.