Skip to main content
9 – 17 UHR +49 8031 3508270 LUITPOLDSTR. 9, 83022 ROSENHEIM
DE / EN
Coding agent Cline Bot Inc. IDE, CLI, Desktop, SDK

Cline

Cline is Cline Bot Inc.'s open-source AI harness for VS Code, JetBrains, terminal and SDK: a coding agent with MCP, checkpoints and free model choice.

Licence Apache-2.0
First release July 2024 (repository created)
Interfaces VS Code, JetBrains, Terminal (CLI), Desktop app, SDK
Platforms macOS, Linux, Windows
Instruction file .clinerules

The seven building blocks in Cline

How Cline implements the building blocks of an AI harness

  1. 1

    Agent loop

    The loop runs in VS Code, JetBrains, the desktop app, or the terminal ('cline'); in the CLI it runs headless via a single prompt with tool auto-approval enabled by default ('--auto-approve', default 'true'), plus optional '--json' streaming for downstream processing.

  2. 2

    Tools and execution environment

    Built-in tools for file editing, running bash commands and browser control can be extended through native MCP servers (managed with 'cline mcp') and through custom tools written with the SDK.

  3. 3

    Context management

    Project rules live in '.clinerules/' or '.cline/rules/', global rules in the Cline Rules directory; 'AGENTS.md', '.cursorrules' and '.windsurfrules' are also recognized.

  4. 4

    State and memory

    Sessions can be inspected and resumed via 'cline history'; team state persisted with '--team-name' and scheduled runs via 'cline schedule' survive restarts and terminal sessions.

  5. 5

    Permissions

    Tool calls are auto-approved by default in the CLI; '--auto-approve false' requires confirmation for each call, and a sandbox mode forced via 'CLINE_SANDBOX' further isolates runs in a dedicated data directory.

  6. 6

    Verification and feedback

    A shadow Git history creates a checkpoint after every tool use; '/undo' and the checkpoint menus can restore files, the task, or both.

  7. 7

    Orchestration

    A coordinator agent can break complex tasks into subtasks and delegate them to specialist agents with their own tools and context; team state persists across sessions via '--team-name'.

Models and model access

Which models the harness uses and how they are obtained

Models and model access
Models No fixed default model - freely selectable across Claude, GPT, Gemini and other model families
Access routes Anthropic, OpenAI, Google Gemini, OpenRouter, AWS Bedrock, Azure/GCP Vertex, Vercel AI Gateway, Ollama/LM Studio, custom OpenAI-compatible endpoints
Free choice of model Yes
Custom endpoint / LLM gateway Yes
A custom OpenAI-compatible endpoint is set in the IDE extension via the 'OpenAI Compatible' provider with a base URL, API key and model ID; in the CLI, setup runs through the same interactive authentication flow ('cline auth') or the global '-P/--provider' and '-k/--key' options - the vendor does not currently document a dedicated command-line option for the base URL.
Features
MCP servers Sub-agents Hooks Skills Plugins / extensions Sandbox Headless / CI SDK IDE integration Checkpoints Resume sessions
Running it with herdr
Supported by herdr Yes
State detection Screen manifest (documented by herdr as less thoroughly tested)
Integration not documented
Per its own documentation, herdr lists Cline as 'detected but less thoroughly tested'; a reliable statement on its integration role is not yet available.
More about herdr →

Enterprise use

Data processing, controls and pricing model according to the vendor

Data processing

When run in your own VPC, per the vendor requests flow directly from your environment to the chosen model provider - Cline itself never sees your code or prompts.

Identity and access

Per the vendor, Cline Enterprise offers SSO and a three-tier role hierarchy (Member/Admin/Owner) with organization-wide permissions.

Admin policies

Central configuration lets you manage model access, MCP controls, rules and workflows for all developers from one dashboard.

Deployment

Besides cloud operation, the vendor names VPC, on-premises and air-gapped deployments, where repositories are, per the documentation, never indexed or cached.

Pricing model

Per the vendor, Enterprise is billed through a credits-based model with usage tracking per user and model; existing provider contracts (BYOK) can alternatively be used with no markup.

Strengths
  • A shared agent core across VS Code, JetBrains, desktop, CLI and SDK gives, per the vendor, identical behaviour on every surface
  • Very broad provider support, including custom OpenAI-compatible endpoints directly via a CLI flag
  • Open-source code under the Apache-2.0 licence
  • Checkpoints via a shadow Git history allow low-risk experimentation
  • Team and scheduling features ('--team-name', 'cline schedule') for recurring and multi-step automation
Limitations
  • Tool calls are auto-approved by default in the CLI - stricter approval requires the explicit '--auto-approve false' option
  • Per herdr's documentation, Cline support there is considered less thoroughly tested than for other harnesses
  • Detailed information on audit logs, retention periods and telemetry is not spelled out on the public enterprise page

Use cases

Typical tasks for this harness

Development with the same agent core across IDE, terminal and desktop
CI/CD automation via 'cline --json' with tool auto-approval enabled by default and NDJSON output for downstream tools
Recurring tasks such as PR summaries via 'cline schedule'
Multi-agent workflows for complex tasks via team mode
Connecting chat surfaces such as Slack or Telegram to agent sessions

Cline is Cline Bot Inc.’s open-source AI harness for VS Code, JetBrains, the terminal, a desktop app, and as an SDK: a coding agent that makes language models operational through tools, MCP servers and an approval system. As of 20 September 2026.

What is Cline?

Cline is developed by Cline Bot Inc. and published in the cline/cline repository under the Apache-2.0 licence; per its GitHub history, the first versions shipped in July 2024. Originally launched as a VS Code extension, the project now describes itself as an “autonomous coding agent as an SDK, IDE extension, or CLI assistant,” and is accordingly available as a VS Code extension, a JetBrains plugin, a native desktop app for macOS and Windows, a terminal CLI, and a Node.js SDK. The CLI is installed via npm install -g cline and started by running cline inside a project directory; the IDE extension is available from the VS Code Marketplace and the JetBrains plugin repository.

Cline as an AI harness

Per the project, all surfaces share the same agent core, so Plan/Act mode, MCP servers, checkpoints, rules and provider configuration behave identically everywhere. In Plan mode, Cline explores the codebase read-only and drafts a strategy; in Act mode it executes it - every file edit and terminal command requires approval per the documentation, unless auto-approve is enabled. In the CLI, tool calls are auto-approved by default; --auto-approve false forces confirmation per call, and a sandbox mode forced via the CLINE_SANDBOX environment variable further isolates runs in a dedicated data directory. Tools are extended through native MCP servers (managed with cline mcp) and through the SDK, which can add custom tools and lifecycle hooks; project rules live in .clinerules/. A coordinator agent can break complex tasks into subtasks and delegate them to specialist agents in team mode (--team-name).

Models and model access

Per its documentation, Cline is not tied to a single provider: supported options include Claude, GPT and Gemini, plus OpenRouter, AWS Bedrock, Azure/GCP Vertex, Vercel AI Gateway, and local models via Ollama or LM Studio. A custom OpenAI-compatible endpoint is set in the IDE extension via the “OpenAI Compatible” provider with a base URL, API key and model ID; in the CLI, setup runs through the same interactive authentication flow (cline auth) or the global -P/--provider and -k/--key options - the vendor does not currently document a dedicated command-line option for the base URL. That means Cline can be run through the AI Gateway, provided the gateway is registered as a custom OpenAI-compatible endpoint.

Enterprise use: data protection and governance

When run in your own VPC, per the vendor requests flow directly from your own environment to the chosen model provider without Cline itself seeing your code or prompts; repositories are, per the documentation, never indexed or cached. Cline Enterprise names SSO, a three-tier role hierarchy, and central configuration for model access, MCP controls and workflows, plus VPC, on-premises and air-gapped deployments. Regardless of the harness used, source code and prompts go to whichever model provider is chosen, and that flow belongs in the record of processing activities beforehand.

Orchestrating Cline with herdr

Per herdr’s own documentation, herdr lists Cline as “detected but less thoroughly tested,” based on a screen manifest; a reliable statement on its integration role is not currently documented. Details and the full list of supported harnesses are on the herdr page.

Where it fits: developer harness and enterprise harness

Cline is a harness for developers working in an IDE, terminal or desktop app - for business users, CompanyGPT is the matching harness, with MCP tools, enterprise knowledge, and roles sourced from Entra ID or Keycloak. The AI Gateway provides the shared cost and governance layer across both. Our overview of all AI harnesses covers further coding agents and orchestrators. We help organizations integrate Cline and comparable harnesses cleanly into existing identity, cost and compliance structures.

From developer harness to enterprise harness

Coding harnesses are tools for developers. Business users, cost control and governance need two more layers.

CompanyGPT – the harness for business users

CompanyGPT gives any language model tools via MCP servers, integrations and n8n, company knowledge via companyRAG, roles from Entra ID or Keycloak, and audit logs – operated in your Azure environment or sovereignly on STACKIT.

Discover CompanyGPT
AI Gateway – cost and governance for every agent

The AI Gateway bundles the model access of your tools and coding agents behind an OpenAI-compatible API: authentication via Entra ID, budgets and cost centres per team, user and agent, guardrails and complete logging.

View the AI Gateway

Frequently Asked Questions

What is Cline?

Cline is an open-source AI harness from Cline Bot Inc. for VS Code, JetBrains, the terminal, desktop, and as an SDK, that turns language models into coding agents through tools, MCP servers and an approval system. The code is published on GitHub under the Apache-2.0 licence.

Is Cline open source?

Yes. The 'cline/cline' repository is licensed under Apache-2.0, which also permits commercial use, modification and redistribution.

Which models does Cline support?

Per its documentation, Cline is not tied to a single provider and supports Claude, GPT and Gemini models among others, plus OpenRouter, AWS Bedrock, Azure/GCP Vertex, Vercel AI Gateway, local models via Ollama or LM Studio, and any OpenAI-compatible endpoint.

Can I connect a custom OpenAI-compatible endpoint?

Yes, via the IDE extension: the 'OpenAI Compatible' provider takes a base URL, API key and model ID. In the CLI, setup runs through the same interactive flow ('cline auth') or the global '-P/--provider' and '-k/--key' options; the documentation does not currently name a dedicated base URL option for the command line.

Can Cline be used in a GDPR-compliant way?

Per the vendor, when run in your own VPC, requests go directly to the chosen model provider without Cline itself seeing your code or prompts. Whether that is GDPR-compliant depends on the model provider chosen and the data processing agreement in place.

Does Cline work with herdr?

Per its own documentation, herdr lists Cline as 'detected but less thoroughly tested', based on a screen manifest; a reliable statement on its integration role is not yet documented.

Introduce coding agents safely in your company

We support selection, model access in the EU, permission design and cost control.