An open-source AI coding extension for VS Code and JetBrains, developed by Continue.dev under the MIT license. Unlike Copilot or Cursor, Continue ships no model of its own. It is a flexible IDE layer that connects to any model you pick.
Claude Opus 4.7, OpenAI frontier models, Gemini 3.1 Pro, Mistral Large, or local models via Ollama, LM Studio, or vLLM all route through the same extension interface.
April 17, 2026: Continue added Claude Opus 4.7 support within 24 hours of Anthropic’s release, alongside Cursor, Windsurf, and Zed.
System Verdict
Pick Continue for maximum model choice with zero vendor lock-in. The extension is free under MIT and stays BYOK. Starter pay-as-you-go at $3 per million tokens covers solo developers who want managed billing and integrations without a seat fee. Team at $20/seat/mo adds shared agents, centralized governance, and $10 of frontier-model credits per seat. Company tier covers custom SSO, BYOK for API keys, and SLAs.
Skip it for out-of-the-box plug-and-play. GitHub Copilot at $10/mo works from day one with less configuration. Cursor at $20/mo bundles models and visual polish.
Who pays what: Free OSS extension stays free with your own keys. Starter $3/M tokens is the smallest paid step for solo devs who want integrations and managed credits. Team $20/seat/mo is right for small teams wanting shared agents and governance. Company custom covers enterprise compliance.
Key Facts
| License | MIT, open source |
| Primary IDEs | VS Code · JetBrains (IntelliJ, PyCharm, WebStorm) · Vim/Neovim |
| Model providers | Anthropic · OpenAI · Google · Mistral · AWS Bedrock · Azure OpenAI · GCP Vertex · Ollama · LM Studio · vLLM |
| Continue Hub | Shared agents, slash commands, rules, MCPs |
| @codebase context | Project indexing with embeddings |
| Custom slash commands | Yes, defined in config |
| Agentic tools | @terminal, @edit, custom functions |
| Free OSS extension | $0, full BYOK use of your own keys |
| Starter pricing | $3 per million tokens, pay-as-you-go, includes Slack/Sentry/Snyk integrations |
| Team pricing | $20/seat/mo with $10/seat frontier-model credits |
| Company pricing | Custom with SSO, BYOK keys, SLAs |
Every data point above was verified against vendor documentation on 2026-05-13. See Sources.
What it actually is
An IDE extension that adds a chat panel, autocomplete, and inline edit commands. Model choice is per-task. A fast local model handles autocomplete; a larger cloud model handles complex reasoning. Both live in the same interface.
@codebase indexes the project with embeddings. The extension answers architecture and usage questions with full file context. Inline editing takes highlighted code plus a natural-language change request.
Continue Hub is the shared layer. Agents, slash commands, rules, and MCP servers publish once and install across team members. Solo developers use it for personal reusable prompts; teams use it for governance and consistency.
Context providers pull documentation, GitHub issues, terminal output, web search, and images. Agentic tools like @terminal and custom functions cover multi-step runs. Config lives in ~/.continue/config.json and is version-controllable.
When to pick Continue
- You want full model choice without IDE lock-in. Swap Opus 4.7 for Mistral Large at will. Run on-device with Ollama when compliance requires it.
- Your team spans VS Code, JetBrains, and Vim. Continue ships on all three; Cursor does not.
- You already have API access through Bedrock, Azure, or Vertex. Route usage through existing enterprise agreements.
- Data residency is non-negotiable. Local Ollama or LM Studio keeps code entirely on-device.
- You want Team governance without losing BYOK. Shared agents and slash commands via Continue Hub.
When to pick something else
- Fastest plug-and-play setup: GitHub Copilot at $10/mo.
- Polished IDE with bundled models: Cursor at $20/mo.
- VS Code autonomous agent with approval gates: Cline. Plan/Act modes, free BYOK.
- CLI workflow with auto git commits: Aider. Free, BYOK, terminal-only.
- Terminal-first agentic runs: Claude Code on Max 5x at $100/mo.
Pricing
Pricing via continue.dev/pricing.
| Plan | Price | Features | Who’s it for |
|---|---|---|---|
| Free OSS extension | $0 | Full open-source extension · BYOK any provider | Individuals on BYOK |
| Starter | $3 per million tokens | Pay-as-you-go input and output · create and run agents · Slack/Sentry/Snyk integrations · purchase frontier-model credits | Pay-as-you-go solo developers |
| Team | $20/seat/mo | $10/seat frontier credits · team agent management · private agent sharing · usage controls · Gmail/GitHub SSO | Small teams |
| Company | Custom | Custom SSO (SAML/OIDC) · BYOK API keys · commitment options · invoicing · SLAs | Compliance-heavy orgs |
Prices verified 2026-05-13 via Continue pricing. The OSS extension stays free and BYOK; model costs are paid directly to the chosen provider, or eliminated entirely with a local Ollama model. Starter is the smallest paid tier and bills $3 per million tokens on input and output.
Against the alternatives
| Continue (Free OSS) | GitHub Copilot | Cursor Pro | |
|---|---|---|---|
| Price | Free + BYOK | $10/mo bundled | $20/mo bundled |
| Model choice | Any BYOK provider | GitHub-hosted models | Mostly bundled |
| IDE support | VS Code · JetBrains · Vim | VS Code · JetBrains · Visual Studio | VS Code fork |
| Setup time | 15-30 min | Under 5 min | Under 10 min |
| Local model option | Yes via Ollama | No | Partial |
| Best viewed as | Flexible BYOK IDE layer | Fastest plug-and-play | Polished bundled IDE |
Failure modes
- No AI included. Must configure and pay for your own model. Adds friction versus Copilot.
- Autocomplete quality tracks the chosen model. A slow or misconfigured local model produces poor completions. Matching model to task takes trial runs.
- Setup overhead. First-time configuration (keys, model selection, context providers, embeddings) runs 15-30 minutes.
- No built-in hosting. Unlike Cursor or Copilot, no Continue-managed model. All infrastructure is user-managed.
- Community-driven support. Documentation is solid. Support runs through GitHub issues and Discord, not dedicated channels.
- Hub governance is new. Team-level policy controls are improving but younger than enterprise-grade competitors.
- Moat is low. Openness and flexibility are the bet. No proprietary model advantage.
- Model provider limits apply. Rate limits, context windows, and quality floors are set by the chosen provider, not by Continue.
Methodology
This page was produced by the aipedia.wiki editorial pipeline, an automated system that ingests vendor documentation, verifies pricing and model details against primary sources, and generates the editorial analysis you are reading. No individual human wrote this review. Scoring follows the four-dimension rubric at /about/scoring/ (Utility × Value × Moat × Longevity, unweighted average). Last verified 2026-05-13 against Continue pricing, Continue docs, and the Continue GitHub repo.
FAQ
Does Continue cost anything? The OSS extension is free under MIT and stays BYOK: you pay your provider directly. Starter at $3 per million tokens is the smallest paid step, with Slack, Sentry, and Snyk integrations. Team at $20/seat/mo adds governance and $10/seat of frontier-model credits. Company tier covers SSO, BYOK keys, and SLAs.
Can I use Continue fully offline? Yes. Configure a local Ollama, LM Studio, or vLLM model and code never leaves the machine. This is one of Continue’s strongest compliance stories.
Continue vs GitHub Copilot, which is better? Copilot is easier on day one and ships quality autocomplete bundled. Continue is free and gives full control over model, provider, and data flow. Teams with existing API contracts or strict data-residency needs usually come out cheaper and more flexible on Continue.
Which IDEs does Continue support? VS Code, JetBrains (IntelliJ, PyCharm, WebStorm), and Vim/Neovim. The extension is near-feature-parity across all three.
What is Continue Hub? A shared registry for agents, slash commands, rules, and MCP servers. Publish once, install across team members. Solos use it for reusable personal prompts; teams use it for consistency and governance.
Sources
- Continue pricing: Starter, Team, Company plans verified 2026-05-13
- Continue docs: configuration, context providers, models
- GitHub: continuedev/continue: source code and releases
- continue.dev: product overview
Related
- Category: AI Coding
- Comparisons: Claude Code vs Continue · Continue vs Cursor · Continue vs Devin · Continue vs GitHub Copilot · Continue vs Val Town
