Cline — An open source AI coding agent for VS Code, terminal and SDK that runs on any model you choose - including local ones.
What it is
Cline is an open source coding agent that edits files, runs commands and verifies results inside VS Code, a terminal CLI, a desktop app or an embeddable SDK, working with cloud or local models under your approval.
Editor's review
Long-form introduction by the BetterPicker editors · checked against the official site · Oct 23, 2026
Cline is the coding agent to pick when you want the agent to follow your stack instead of a vendor's. It is an open source agent that ships as a VS Code extension, a terminal CLI, a desktop app and an embeddable SDK, and it runs on Claude, GPT, Gemini, local Ollama or LM Studio models, or any OpenAI-compatible endpoint - bring your own key or buy inference at cost. The tool itself is free for individual developers; usage-based model costs are the only bill. Reviewers describe real end-to-end wins - a working app generated from one prompt, files edited line by line, a browser opened to verify the result - balanced by an honest caveat: small local models stall on hard tasks, and someone still has to approve the steps.
What it does well
Model choice is genuine: Claude, GPT, Gemini, OpenRouter, AWS Bedrock, DeepSeek and local Ollama or LM Studio models all plug into the same workflow, and users report swapping providers when prices, quotas or privacy needs change without changing tools.
Plan, then act, under your control: Plan mode aligns on a strategy before Act mode executes, every edit arrives as a diff with checkpoints and one-click undo, and shell commands stream their output live; reviewers single out the per-request cost readout as unusually transparent.
Extends past the editor: MCP servers and a marketplace bolt on databases, APIs and infrastructure, .clinerules files carry team conventions inside the repo, and the same agent runs headlessly in CI or answers messages from Slack.
Who it's for
Cline fits developers who want one agent across editor, terminal and pipelines without marrying a model vendor. It is especially compelling if you already hold API keys, want local models for private codebases, or need to embed the runtime in your own product through the SDK. The visible diffs and per-step approval also make it approachable for cautious adopters. Teams that need SSO, role-based access and centralized billing move to the paid Enterprise tier. If you would rather have a fully managed subscription where model, hosting and interface arrive bundled, a closed assistant will feel simpler - but it will also decide your stack for you.
Where it falls short
Output tracks the model you attach: walkthroughs repeatedly show small local models, 8B-class ones especially, failing mid-task on complex changes; the smooth experience reviewers praise generally assumes premium models behind the meter.
Costs scale with autonomy: pay-per-inference pricing is transparent but open-ended, and users report watching credits burn through long agentic sessions; budgeting discipline is on you, not the tool.
Supervision is part of the job: an agent that edits files, runs commands and drives a browser needs review; auto-approve is a convenience that trades away exactly the checking that keeps a run safe, and enterprise controls wait in the paid tier.
Specs at a glance
Facts from the official site · not editorial opinion
| Price | Free for individuals; usage-based AI inference or BYOK; Enterprise custom |
|---|---|
| Surfaces | VS Code extension, terminal CLI, desktop app, embeddable SDK |
| Open source | Yes |
| Models | Claude, GPT, Gemini, DeepSeek, OpenRouter, Bedrock, Vertex; local via Ollama and LM Studio |
| Core capability | Autonomous coding agent that edits files, runs commands and verifies results with step approval |
| Integrations | MCP servers and marketplace; .clinerules project rules; Slack, Linear and CI hooks |
| Best for | Developers who want one agent across editor and terminal without model vendor lock-in |
Frequently asked questions
▸Is Cline free to use?
The open source extension, CLI and desktop app are free for individual developers. You pay only for AI inference - either through the Cline provider at cost or by connecting your own API keys; there are no seat fees or subscriptions on the open source version.
▸Which AI models does it support?
Claude, GPT, Gemini and DeepSeek through major providers, plus OpenRouter, AWS Bedrock, GCP Vertex and any OpenAI-compatible endpoint. Local models via Ollama or LM Studio work too, though reviewers note that small local models struggle with complex tasks.
▸Can I use it outside VS Code?
Yes. The same agent runtime ships as a terminal CLI, a native desktop app that needs no editor, and an SDK for embedding in your own products; JetBrains support is part of the Enterprise tier.
▸When does an organization need the Enterprise plan?
When it needs SSO, role-based access control, centralized billing, audit logs, SLAs or the JetBrains extension. Individual developers and most small teams can stay on the free, open source setup indefinitely.
Video reviews
3 review videos aggregated · praise and criticism included alike · click through to the original video
Channels that covered it
Channels are aggregated as sources only — we don’t rate creators
Related tools
Where to go next
External links open in a new tab; external content is independent of this site.
Link down? Every object page is re-checked monthly.