Continue
Continue is an open-source AI code assistant that installs as an extension in VS Code and JetBrains IDEs, providing autocomplete, chat, and edit capabilities with full control over which AI models and context providers are used — serving as the open-source alternative to GitHub Copilot where developers can bring their own models (OpenAI, Anthropic, local Ollama models), customize prompts and workflows, and maintain complete transparency over how their code is processed.
What Is Continue?
- Definition: An open-source IDE extension (Apache 2.0 license) that provides AI-powered autocomplete, conversational chat about code, and inline edit capabilities — with the critical distinction that users choose and configure their own AI models rather than being locked into a single provider.
- Bring Your Own Model: Unlike Copilot (locked to GitHub/OpenAI), Continue supports any LLM provider — OpenAI (GPT-4), Anthropic (Claude), Google (Gemini), local models via Ollama, or any OpenAI-compatible API endpoint.
- Full Customization: Custom system prompts, context providers (add documentation, wiki pages, or database schemas to AI context), and slash commands — define workflows that match your team's specific practices.
- IDE Support: Available for VS Code and JetBrains (IntelliJ, PyCharm, WebStorm, etc.) — covering the two dominant IDE ecosystems.
Key Features
- Autocomplete (Tab): Inline code suggestions as you type — similar to Copilot's ghost text, powered by your chosen model. Supports FIM (Fill-in-the-Middle) models for context-aware completions.
- Chat (Cmd+L): Conversational AI panel in the IDE — ask questions about your codebase, get explanations, discuss architecture decisions. Supports adding files and folders to context.
- Edit (Cmd+I): Select code, describe changes ("Refactor this to use async/await"), and the AI modifies the selection in-place with a diff preview.
- Context Providers: Extensible system for adding context to AI conversations — `@file` (specific files), `@folder` (directory contents), `@docs` (documentation URLs), `@codebase` (semantic search), `@terminal` (recent terminal output).
- Slash Commands: Custom commands like `/test` (generate tests), `/doc` (generate documentation), `/fix` (fix errors) — configurable per project.
Continue vs. Alternatives
| Feature | Continue | GitHub Copilot | Cursor | Tabnine |
|---|---|---|---|---|
| License | Open-source (Apache 2.0) | Proprietary | Proprietary | Proprietary |
| Model Choice | Any (BYO) | GPT-4o (fixed) | Multiple (configurable) | Cloud or local |
| Customization | Full (prompts, context, commands) | Limited | Moderate | Limited |
| IDE Support | VS Code + JetBrains | VS Code + JetBrains + more | VS Code fork only | All major IDEs |
| Cost | Free (+ model API costs) | $10-39/month | $20/month | $12/month | |
| Data Privacy | Full control (self-host models) | Code sent to GitHub/OpenAI | Code sent to Cursor | Local option available |
Configuration Example
Continue is configured via a JSON file (`.continue/config.json`) in your project:
- Models: Define which models handle chat, autocomplete, and edits separately — use fast local models for autocomplete and powerful cloud models for complex chat.
- Context Providers: Configure documentation sources, database schemas, or custom APIs as context providers.
- Custom Slash Commands: Define project-specific commands that inject templates, run scripts, or perform specialized transformations.
Continue is the open-source AI code assistant that gives developers full control over their AI coding experience — combining Copilot-like autocomplete with chat and edit capabilities while letting teams choose their own models, customize prompts, and maintain complete transparency and data sovereignty over how their code is processed.