Continue

Continue is an open-source AI coding assistant that runs inside VS Code and JetBrains IDEs. The distinguishing feature is control: you configure exactly which models power chat, autocomplete, and edit — Claude, GPT, DeepSeek, local Ollama models, or anything in between — and you can mix providers freely (fast local model for autocomplete, frontier model for chat). Continue supports the Model Context Protocol as a first-class client, so any MCP server (GitHub, Postgres, Sentry, Notion, GitMCP) becomes a set of tools your assistant can call. Continue Hub is a marketplace where teams publish reusable blocks: assistants (pre-configured chat personalities), models, prompts, rules, docs, and MCP servers. A team can standardize its AI setup by publishing an assistant to Hub and having everyone install it, so every dev has the same prompts, rules, and connected tools. Continue's agent mode plans multi-step edits across a codebase and asks for confirmation before writing. Because the extension is open source and BYO-model, teams that need to run everything self-hosted or with a specific model policy can do so. Pricing: the extension is free forever (open source, MIT-licensed); Continue Hub has a Solo tier free with limited private blocks, Teams at $10/user/month (private blocks and analytics), and Enterprise custom with SSO, on-prem hub, and audit logs. Continue fits engineering teams that want AI-assisted coding but keep control of models, tools, and prompts.

What Continue does in practice

Open-source assistant for VS Code and JetBrains. Bring-your-own-model: Anthropic, OpenAI, Ollama, local, any provider. First-class MCP client — any MCP server becomes callable tools. Continue Hub for sharing assistants, rules, prompts across a team. Agent mode for multi-step edits with confirmation. Works fully self-hosted with local models for privacy-sensitive teams. These capabilities should be judged as a workflow, not isolated checkboxes. Concrete best-fit scenarios include: Teams that need to standardize AI coding setup across many devs. Regulated environments requiring self-hosted models. Developers who want a specific model per task (fast local + smart cloud). Building custom agent tools via MCP servers plugged into the IDE.

Pricing and total cost

Open-source extension is listed at Free; VS Code and JetBrains extensions are MIT-licensed and free forever.. Hub — Solo is listed at Free; Publish and consume public blocks; limited private blocks.. Hub — Teams is listed at $10/user/month; Private team blocks, shared assistants, analytics.. Hub — Enterprise is listed at Custom; SSO, on-prem hub, audit logs, dedicated support.. These figures come from the staging record and vendor pages were fetched during this revision, but pricing is mutable. Recheck the live pricing page, checkout, or quote before publishing or buying. Include required seats, usage credits, model or API charges, storage, implementation, support, taxes, and overages. A free tier can prove compatibility; it does not necessarily model production volume or support needs.

Strengths and tradeoffs

Practical strengths include: Open-source assistant for VS Code and JetBrains. This directly supports the primary workflow. Bring-your-own-model: Anthropic, OpenAI, Ollama, local, any provider. First-class MCP client — any MCP server becomes callable tools. Continue Hub for sharing assistants, rules, prompts across a team. Important constraints include: At least one tier requires a sales quote, preventing a complete public budget calculation. Features and limits can change by plan, billing interval, region, and release; verify current documentation. MCP connectivity expands the permission surface and requires narrow scopes, approvals, logs, and credential rotation. A representative pilot is necessary because a feature list does not establish reliability or operational fit. A capable product can still be the wrong operational choice. Document who reviews outputs, who owns failures, what data may enter the system, and how work continues during an outage.

How to evaluate Continue

Use a two-week proof of concept on a noncritical repository or workflow. Define four repeatable tasks: one small change, one multi-file change, one failure-recovery scenario, and one permissions-sensitive action. Measure setup time, completion rate, human review minutes, retries, API or token cost, and defects found after acceptance. Pin versions, inspect generated changes, and test rollback. Give the tool only the repositories, APIs, and scopes needed for the trial. Verify export, backup, logging, rate-limit, and offboarding behavior before treating a fast demo as production readiness.

Bottom line

Continue is worth shortlisting when its capabilities match a measured workflow and the buyer can verify current limits. Do not select it from a feature list alone. Compare at least two alternatives using the same inputs, acceptance criteria, and cost model. Preserve test results so renewal is based on outcomes rather than adoption anecdotes.