Skip to main content

AI Coding Agent

Cline

Open-source coding agent that plans changes, then edits files and runs terminal commands with your approval. It runs from VS Code, JetBrains IDEs, a CLI, or a desktop app, on the model provider you choose.

The 30-Second Take

Pick Cline if you want to keep your editors and run an agent on your own model provider for free.

It reads the codebase, edits files and runs commands with your approval, in VS Code, JetBrains or the terminal. Bedrock, Vertex or a self-hosted endpoint all work, on your own key.

Best fit
Teams that keep their current IDEs, pay their own model provider, and want engineers to approve each step.
Main constraint
Cost and output quality follow the model and provider you configure, and Enterprise controls stop at the editor extensions.
First move
Start every task in Plan mode, limit auto-approve to reads, and keep checkpoints on.

Why

Watch the model you configure, because cost and output quality follow it. Enterprise's provider, MCP and telemetry controls reach only the two editor extensions, so turn telemetry off by hand in the CLI.

Practitioner sentiment, all practitioners+39 Positive

70 practitioners, 112 sources reviewed, Updated Nov 1, 2026How this is measured

Read the GILD Review
Save a product your organization uses.Keep up with Cline without adding it to your stack.Visit Website
Operating stack
Agentic Software Development

Comparisons

How does Cline compare with the alternatives?

See all 6 comparisons

  1. Cline vs Cursor

    • AI Coding Agents

    Pick Cline to add an agent to the editors engineers already use, Cursor included, on the provider your team pays directly. Pick Cursor to move the team onto one editor with seats and model usage sold together.

  2. Cline vs GitHub Copilot

    • AI Coding Agents

    Pick GitHub Copilot if model spend should run through GitHub seats with included AI credits and a cloud agent that opens pull requests. Pick Cline if it should run through your own provider account, with no seat fee.

  3. Cline vs Claude Code

    • AI Coding Agents

    Both edit files and run commands from the terminal and from VS Code and JetBrains IDEs. Claude Code runs Claude models through a Claude subscription, an Anthropic Console account, or a supported cloud provider. Cline runs on any provider your team configures, a Claude subscription included.

Operator Insights

What are engineering leaders learning from deploying Cline?

Buying and deployment

What engineering leaders should verify before a broader rollout.

Sourced from Cline documentation

How you buy it

US$9.99

A month for ClinePass. The extension, CLI, and desktop app are free with your own key.

Source 1, Cline Bot Inc., Cline pricing: Cline pricingSource 9, Cline Bot Inc., Cline Desktop docs: Cline Desktop docsSource 2, Cline Bot Inc., Authorization and bring your own key: Authorization and bring your own keySource 5, cline.bot, ClinePass: ClinePass

Details

The open-source extension and CLI are free for individual developers, and the desktop app is open source. Model access is your own provider key, pay-as-you-go Cline credits, or ClinePass at US$9.99 per month standard rate; Enterprise is custom-priced through sales.

Cost drivers

Model usage

Your provider's rates with your own key, or Cline credits pay-as-you-go.

Source 1, Cline Bot Inc., Cline pricing: Cline pricingSource 2, Cline Bot Inc., Authorization and bring your own key: Authorization and bring your own keySource 6, Cline Bot Inc., ClinePass docs: ClinePass docs

Details

Model usage is the cost. With your own key you pay the provider's rates directly; Cline credits are used pay-as-you-go on paid models; ClinePass quotas reset on 5-hour, weekly, and monthly windows.

Where it runs

Editor, CLI, desktop

VS Code and its forks, JetBrains, a terminal CLI, a desktop app, and cloud agents.

Source 7, docs.cline.bot, Installing Cline: Installing ClineSource 8, cline.bot, Cline for Desktop: Cline for DesktopSource 10, Cline Bot Inc., Cloud agents: Cloud agents

Details

Extension for VS Code (also Cursor, Windsurf, and VSCodium) and JetBrains IDEs, a terminal CLI, a desktop app in beta for macOS, Windows, and Linux, and cloud agents in per-task sandboxes.

What to verify

Telemetry on

On by default. Organization telemetry controls apply in the VS Code and JetBrains extensions only.

Source 4, Cline Bot Inc., Telemetry: TelemetrySource 3, Cline Bot Inc., Enterprise provider configuration: Enterprise provider configuration

Details

Telemetry is on by default, and prompts reach Cline only on its own providers. Organization provider, MCP and telemetry controls apply in the VS Code and JetBrains extensions only; the CLI receives only organization rules and workflows.

Signals

What is the current evidence saying about Cline?

A current read of what developers, teams, and technical leaders report publicly about Cline, with every finding linked to its sources.

Whose sources count

Practitioner sentiment, all practitioners

+39 Positive net

of 73 sources that take a side, each author counted once.

How this figure is made

18 more state no overall view and cast no vote.

Leaders and teams have 8 accounts; GILD prints their sentiment from 20.

Current sentiment

Positive, mixed and negative sources

The score counts them as 40 positive, 17 mixed and 13 negative votes, one per author.

Across public accounts the read is positive. Model and provider choice is the most praised capability. Spending tokens on your own API key is the most frequent concern, followed by update regressions and results that depend on the model.

Use a task pilot to measure accepted work and review time.

Evidence field. All 112 sources, one mark each.

Arrange by

Each mark is one source: positive, mixed, negative, no opinion stated, or documentation (not counted).

By sentiment, All practitioners. Every source GILD read, grouped as the field draws it.
GroupSourcesPositiveMixedNegativeNo opinion stated and not counted in the scoreDocumentation, checked and not counted
Positive43430000
Mixed17017000
Negative13001300
Not counted21000210
Documentation18000018
Every source GILD read, by the quarter it was published, as the field's time axis counts it. Documentation is dated by the day it was checked.
QuarterSources
Q1 20258
Q2 202523
Q3 202526
Q4 20256
Q1 20267
Q2 202610
Q3 202611
Q4 20263
Documentation, checked18

By sentiment

18 documentation pages check facts and are not counted.

Hover a mark for a quick look, click it or press Enter to open its source.Tap a mark to open its source, then step to its neighbors with Previous and Next.

What practitioners praise

Top praise of 4. Each mark is a source on it; filled marks praise it.

  1. Model and Provider ChoiceAll 20 sources on it praise it

Most common concerns

Top concern of 5. Each mark is a source on it; filled marks raise it.

  1. Token Spend on Your Own Key17 of 29 sources on it raise it as a concern
All 112 sources behind these findings
  1. Broken Anthropic prompt caching on tool calls in v4.1.25 ("Next" engine) resulting in cost spikes.nicolaspe, GitHub user (self-described: VS Code extension on Windows with an Anthropic API key), GitHub issue on cline/cline, published Oct 24, 2026
  2. Agent silently set CLINE_CAPTURE_PROVIDER_REQUEST=full → Plaintext credentials on disk → Unauthorized paid calls on an unrelated provider keyIRwDONE, GitHub user (self-described: Cline Desktop beta on Windows, daily sessions in Act mode), GitHub issue on cline/cline, published Oct 19, 2026
  3. Cline pricingCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  4. AuthorizationCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  5. Cline (usage-billing)Cline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  6. ClinePassCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  7. Other 30+ providersCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  8. Plan & Act ModeCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  9. Auto Approve & YOLO ModeCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  10. CheckpointsCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  11. RulesCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  12. MCPCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  13. CLI overviewCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  14. Cline DesktopCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  15. Cloud agentsCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  16. Cline EnterpriseCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  17. Enterprise provider configurationCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  18. MCP server controlsCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  19. Cline telemetryCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  20. SSO setupCline Bot Inc. (Official), Product documentation, checked Oct 12, 2026
  21. CRITICAL: run_commands on Windows (PowerShell) - mangled quoting led to rmdir /s /q running against a drive root, mass deletion of user dataKorn94, Cline user on Windows (self-described: VS Code on Windows, Act mode with run_commands), GitHub issue, published Oct 6, 2026
  22. [Desktop App] Sessions marked failed keep consuming for 20+ hours; FREE-labeled models bill credits; ~183k input tokens per callromemozu, ClinePass and Cline Desktop user (self-described: Cline Desktop with ClinePass and Cline credits), GitHub issue, published Sep 18, 2026
  23. Early access to the Cline open-source Desktop app (X post)Shivay Lamba, AI architect, Docker Captain and CNCF Ambassador (self-described: Solo test of Cline Desktop on coding, scheduled reviews and handoffs), X post, published Sep 14, 2026
  24. Comment on: Which tools do Claude, Codex and Cursor choose? We measured 17k runs to find outNeywiny, Hacker News commenter, self-described, Hacker News comment, published Sep 4, 2026
  25. Cline on Autopilot: Why I Put Limits on My AgentJuan Torchia, software architect (self-described: Codebases with history), Personal blog post, DEV Community, published Sep 2, 2026
  26. Cline truncates tool output & file content in context (forces splitting large files into chunks)rvnlord, Cline user filing a GitHub issue (self-described: C# .NET developer), GitHub issue, published Aug 14, 2026
  27. Plan mode not respecting edit permissions + degraded task quality after v4.x SDK migrationumair1808, Cline user filing a GitHub issue (self-described: DeepSeek through a self-hosted LiteLLM proxy), GitHub issue, published Aug 10, 2026
  28. Why the Cline Extension for VS Code Fails as an "Autonomous Agent" Under LinuxStanislav Nabatnikov, developer (self-described: Real project on Kubuntu Linux), Personal blog post, DEV Community, published Jul 11, 2026
  29. Cline hangs completely after skipping a command execution, requires VS Code restartstefanCCS, Cline user filing a GitHub issue, self-described, GitHub issue, published Jul 9, 2026
  30. GLM-5.2: Considerations for enterprise teams starting out with open-weight modelsRobat Williams, developer building client web applications at Scott Logic (Scott Logic engineering team running A/B trials of AI coding setups on open-source issues), Company blog post, Scott Logic Blog, published Jul 8, 2026
  31. Git commit --amend should not be considered a "safe command"BroadlyWhitaker, Cline user filing a GitHub issue, self-described, GitHub issue, published Jul 7, 2026
  32. Comment on: Fable 5 is BackMarha01, Hacker News commenter (self-described: Opus for Plan and Sonnet for Act), Hacker News comment, published Jul 1, 2026
  33. Cline 4.0.0 issue: MCP server settings no longer read after upgradinglewisl, Cline user filing a GitHub issue (self-described: Fresh install of Cline 4.0.0 with an MCP server), GitHub issue, published Jun 27, 2026
  34. Comment on: CTOs Agree: Cognitive Debt Is the New Technical Debtfoobarchu, Hacker News commenter (self-described: Spring Boot 3 to 4 migration on Claude), Hacker News comment, published Jun 21, 2026
  35. Comment on: Ask HN: What are tools you have made for yourself since the advent of AI?TrueBlueKiwi, Hacker News commenter (self-described: Web app on Claude Sonnet 4.7), Hacker News comment, published Jun 13, 2026
  36. Comment on: an Ask HN thread on the moment generative AI first surprised peoplecdelsolar, Hacker News commenter (self-described: Works on voice AI at Salesforce), Hacker News comment, published Jun 6, 2026
  37. Comment on: DeepSeek makes the V4 Pro price discount permanentfirecall, Hacker News commenter (self-described: VS Code with DeepSeek V4, in place of GitHub Copilot), Hacker News comment, published May 23, 2026
  38. [Bug]: Local Ollama models (Qwen 2.5 Coder) trapped in infinite loop emitting raw JSON instead of triggering tools due to strict XML parser constraintLittleRainC, GitHub user (self-described: Cline 3.83.0 in VS Code on a local Ollama model), GitHub issue on cline/cline, published May 18, 2026
  39. Comment on: Local AI needs to be the normstuaxo, Hacker News commenter (self-described: Local models), Hacker News comment, published May 11, 2026
  40. Comment on: Vibe coding and agentic engineering are getting closer than I'd likerglover, Hacker News commenter, self-described, Hacker News comment, published May 6, 2026
  41. Comment on: AI's economics don't make sensechankstein38, Hacker News commenter (self-described: Simple project on a US$15 API balance), Hacker News comment, published Apr 28, 2026
  42. Comment on: Cursor 3zenoprax, Hacker News commenter (self-described: VSCodium with OpenRouter, after Cursor), Hacker News comment, published Apr 2, 2026
  43. Comment on: 1M context is now generally available for Opus 4.6 and Sonnet 4.6a_e_k, Hacker News commenter (self-described: Developer at a company with an enterprise model plan), Hacker News comment, published Mar 14, 2026
  44. Clinejection: Compromising Cline's Production Releases just by Prompting an Issue TriagerAdnan Khan, security researcher (self-described: Independent security research on Cline's release pipeline), Security research post, adnanthekhan.com, published Feb 9, 2026
  45. Agressive context compaction with opus-4.6:1mkhkiley, Cline user filing a GitHub issue (self-described: Long tasks on Claude Opus 4.6 with a 1M-token window), GitHub issue, published Feb 8, 2026
  46. Comment on: Coding agents have replaced every framework I usedrglover, Hacker News commenter (self-described: Daily use for nearly two years on Sonnet and Opus), Hacker News comment, published Feb 7, 2026
  47. AUMOVIO Boosts Software Development with an Agentic Coding Assistant Powered by Amazon BedrockAUMOVIO and AWS, Martin Kraus and Amir Mahdi Namazi of AUMOVIO, with AWS authors (self-described: AUMOVIO high-performance computer software teams, hackathon and pilot build with AWS), Vendor case study, AWS for Industries blog, published Jan 15, 2026
  48. Comment on: AI coding assistants are getting worse?wickedsight, Hacker News commenter (self-described: Work environment without cloud AI), Hacker News comment, published Jan 9, 2026
  49. Comment on: I came back from Cursor to VS Codemasto, Hacker News commenter (self-described: VS Code, pay-as-you-go for bursty use), Hacker News comment, published Jan 5, 2026
  50. Comment on: A guide to local coding modelsDrAwdeOccarim, Hacker News commenter (self-described: Local Devstral Small 2 served by vLLM on a DGX Spark), Hacker News comment, published Dec 22, 2025
  51. Comment on: A guide to local coding modelsjszymborski, Hacker News commenter (self-described: Local quantized models), Hacker News comment, published Dec 22, 2025
  52. Comment on: GPT-5.2tgtweak, Hacker News commenter (self-described: Long project on Claude with a 1M-token context), Hacker News comment, published Dec 12, 2025
  53. Comment on: Everyone in Seattle hates AIsparrc, Hacker News commenter at AWS (self-described: AWS, Seattle), Hacker News comment, published Dec 3, 2025
  54. Comment on: Gemini 3syspec, Hacker News commenter (self-described: Developer with work-provided model access), Hacker News comment, published Nov 18, 2025
  55. Comment on: Claude Haiku 4.5Fergusonb, Hacker News commenter, self-described, Hacker News comment, published Oct 16, 2025
  56. Comment on: Claude Code 2.0esperent, Hacker News commenter (self-described: Compared with Claude Code, switching APIs by price), Hacker News comment, published Sep 30, 2025
  57. Comment on: Gentoo AI Policyjjmarr, Hacker News commenter (self-described: Software engineer at AMD), Hacker News comment, published Sep 18, 2025
  58. Comment on: Vibe coding as a coding veteran: from 8-bit assembly to English-as-code_fat_santa, Hacker News commenter, self-described, Hacker News comment, published Sep 1, 2025
  59. "Cline uses complex prompts and iterative task execution that may be challenging for less capable models..."John-Codes, Cline user filing a GitHub issue (self-described: Qwen3 Coder), GitHub issue, published Aug 31, 2025
  60. Cline: Vulnerable To Data Exfiltration And How To Protect Your Datawunderwuzzi, security researcher (self-described: Security testing of Cline), Security research post, Embrace The Red, published Aug 27, 2025
  61. Research, Review, RebuildRahul Ramesh, Consultant Developer at Thoughtworks (Thoughtworks modernization of the Bahmni hospital system front end from AngularJS to React and FHIR), Engineering article, martinfowler.com, published Aug 27, 2025
  62. Automatic chat compacting at 160k tokens loses critical context and causes Cline to restart completed workmosmondor, Cline user filing a GitHub issue (self-described: Long tasks on Claude Sonnet 4 via OpenRouter), GitHub issue, published Aug 24, 2025
  63. Comment on: AWS CEO says using AI to replace junior staff is 'Dumbest thing I've ever heard'ramesh31, Hacker News commenter (self-described: Developer at a company), Hacker News comment, published Aug 22, 2025
  64. AI isn't just a coding partner, it can be a deployment partner, tooDharanikumar Premkumar, Tech Lead and Application Architect at Thoughtworks (Thoughtworks team deploying a five-service event-driven platform on AWS for an automotive client), Company blog post, Thoughtworks Insights, published Aug 21, 2025
  65. Comment on: GPT-5 for Developerst1amat, Hacker News commenter who provisions tools for a team (self-described: Team lead choosing tools for the team), Hacker News comment, published Aug 8, 2025
  66. Comment on: I want everything local - Building my offline AI workspaceramesh31, Hacker News commenter (self-described: Qwen and GLM models through OpenRouter), Hacker News comment, published Aug 8, 2025
  67. Comment on: a quote-reply extension for Slack threads, built with GPT-5KeysToHeaven, Hacker News commenter (self-described: Chrome extension built with GPT-5), Hacker News comment, published Aug 8, 2025
  68. Comment on: Claude Opus 4.1chisleu, Hacker News commenter (self-described: Opus or Gemini 2.5 Pro for Plan, Sonnet for Act), Hacker News comment, published Aug 6, 2025
  69. Comment on: ScreenCoder: An intelligent UI-to-code generation system_fat_santa, Hacker News commenter (self-described: UI components built with a Figma MCP server), Hacker News comment, published Aug 6, 2025
  70. Comment on: Open models by OpenAINicoJuicy, Hacker News commenter (self-described: Local gpt-oss 20B on an RTX 3090 through Ollama), Hacker News comment, published Aug 5, 2025
  71. Comment on: Tokens are getting more expensivereilly3000, Hacker News commenter (self-described: OpenRouter, choosing models per task by price), Hacker News comment, published Aug 4, 2025
  72. AI evals for MCP in AIOpsZichuan Xiong, Global Head of AIOps at Thoughtworks (Thoughtworks Managed Services AIOps build for a Latin American bank), Company blog post, Thoughtworks Insights, published Jul 31, 2025
  73. LISP Interpreter: Created by Cline and Qwen 3 Coder for less than US$1Eleanor Berger, developer (self-described: Solo experiment, Rust), Project write-up, GitHub Gist (intellectronica), published Jul 24, 2025
  74. Cline autonomously delete files without keeping track of the deleted/changed files. Very Dangerous and Critical Issue!!!Revenoti, Cline user filing a GitHub issue, self-described, GitHub issue, published Jul 23, 2025
  75. How a Real AI Agent Works Under the HoodLorenz Hofmann-Wellenhof, engineer (self-described: Solo; a couple of months of daily coding with Cline next to Cursor), Newsletter post, The Growing Engineer (Substack), published Jul 22, 2025
  76. cline feels like an upgradeVincent D. Warmerdam, data scientist and developer (self-described: Personal use, trying a new tool each week), Personal blog post, koaning.io, published Jul 21, 2025
  77. Comment on: Cognition (Devin AI) to Acquire Windsurflizardking, Hacker News commenter, self-described, Hacker News comment, published Jul 15, 2025
  78. Comment on: Kiro: A new agentic IDEtheturtletalks, Hacker News commenter (self-described: Cline and Roo Code on OpenRouter, then Claude Code on the Max plan), Hacker News comment, published Jul 14, 2025
  79. Comment on: OpenAI's Windsurf deal is off, and Windsurf's CEO is going to Googlemilofeynman, Hacker News commenter, self-described, Hacker News comment, published Jul 12, 2025
  80. Comment on: OpenAI's Windsurf deal is off, and Windsurf's CEO is going to Googleefitz, Hacker News commenter, self-described, Hacker News comment, published Jul 11, 2025
  81. Many Unexpected API Response still burning creditsandreas83, Cline user filing a GitHub issue (self-described: Grok 3 and Gemini 2.5 Pro on paid credits), GitHub issue, published Jul 9, 2025
  82. Case Study + Deep Dive: Telemedicine Support Agents with LangGraph/MCPDan Mason, AI product and technology leader at Stride (Stride team of two engineers and a designer building a patient-support agent for a telemedicine client), Conference talk with transcript, AI Engineer World's Fair 2025, published Jun 22, 2025
  83. Comment on: Claude Code feels like magic because it is iterativeaussieguy1234, Hacker News commenter, self-described, Hacker News comment, published Jun 19, 2025
  84. Comment on: Generative AI coding tools and agents do not work for meqinsig, Hacker News commenter, self-described, Hacker News comment, published Jun 17, 2025
  85. [Improvement] Cline Does Not Treat eval as an unsafe command, allowing execution of Arbitrary Shell Commands, tighten permissionsgetsec, for a team building an MCP server on internal APIs (self-described: Team building an MCP server for internal APIs, Cline with Gemini 2.5 Flash), GitHub issue, published Jun 6, 2025
  86. Comment on: Why Cline doesn't index your codebaseNitpickLawyer, Hacker News commenter (self-described: Local models such as Devstral), Hacker News comment, published May 27, 2025
  87. Comment on: Claude 4jjmarr, Hacker News commenter (self-described: Junior developer contributing to LLVM), Hacker News comment, published May 22, 2025
  88. Comment on: Show HN: Vibe coding from your phonechrisnolet, Hacker News commenter (self-described: Founder of Pocket, works on Vision Pro at Apple), Hacker News comment, published May 20, 2025
  89. Comment on: A flat pricing subscription for Claude Codebicepjai, Hacker News commenter (self-described: Solo, Rust side project), Hacker News comment, published May 15, 2025
  90. Comment on: Perverse incentives of vibe codingtheshrike79, Hacker News commenter (self-described: OpenRouter, Gemini 2.5 for Plan and Claude 3.7 for Act), Hacker News comment, published May 15, 2025
  91. Comment on: The unreasonable effectiveness of an LLM agent loop with tool usecodethief, Hacker News commenter (self-described: Solo, Android prototype), Hacker News comment, published May 15, 2025
  92. Comment on: Ask HN: Cursor or Windsurf?int_19h, Hacker News commenter, self-described, Hacker News comment, published May 12, 2025
  93. Comment on: Ask HN: Cursor or Windsurf?shmoogy, Hacker News commenter, self-described, Hacker News comment, published May 12, 2025
  94. Comment on: A flat pricing subscription for Claude Codeasaddhamani, Hacker News commenter (self-described: Compared with Cursor and Copilot, on a personal API key), Hacker News comment, published May 9, 2025
  95. Comment on: Ask HN: How much better are AI IDEs vs. copy pasting into chat apps?Flux159, Hacker News commenter (self-described: Web app, production code changes), Hacker News comment, published May 8, 2025
  96. Comment on: Ask HN: How much better are AI IDEs vs. copy pasting into chat apps?neilsharma, Hacker News commenter (self-described: Frontend React and TypeScript), Hacker News comment, published May 8, 2025
  97. How Cline Quietly Changed The Game For Code CopilotsJos van der Westhuizen, co-founder and CEO of Coplay (self-described: Coplay team, 60,000-line project), Company blog post, Coplay blog, published May 6, 2025
  98. Comment on: OpenAI o3 and o4-minichaboud, Hacker News commenter (self-described: Individual developer with AWS Bedrock access), Hacker News comment, published Apr 17, 2025
  99. Comment on: OpenAI Codex CLI: Lightweight coding agent that runs in your terminalrazemio, Hacker News commenter (self-described: Compared with Aider, on Claude Sonnet and Gemini 2.5 Pro), Hacker News comment, published Apr 17, 2025
  100. Comment on: OpenAI looked at buying Cursor creator before turning to WindsurfCalmStorm, Hacker News commenter (self-described: Compared with Cursor, on DeepSeek V3), Hacker News comment, published Apr 17, 2025
  101. Comment on: Roo or Cline? We're building a supersetdanielbln, Hacker News commenter, self-described, Hacker News comment, published Apr 12, 2025
  102. Comment on: Claude's Max PlanSkyPuncher, Hacker News commenter (self-described: Full-time software developer), Hacker News comment, published Apr 9, 2025
  103. Cline (Technology Radar, Trial)Thoughtworks Technology Advisory Board (self-described: Technology Radar blip, April 2025 edition), Technology Radar assessment, Thoughtworks Technology Radar, published Apr 2, 2025
  104. Comment on: Ask HN: What's the best way to get started with LLM-assisted programing?timbritt, Hacker News commenter (self-described: Solo, building applications), Hacker News comment, published Apr 1, 2025
  105. Comment on: Ask HN: What's the best way to get started with LLM-assisted programing?inheritedwisdom, Hacker News commenter speaking for their team (self-described: Data engineering team on Databricks, early rollout), Hacker News comment, published Mar 31, 2025
  106. The role of developer skills in agentic codingBirgitta Böckeler, Distinguished Engineer and AI-assisted delivery expert at Thoughtworks (self-described: Changing existing codebases with Cursor, Windsurf and Cline), Engineering memo, martinfowler.com, published Mar 25, 2025
  107. Building My First Web App with the Help of AIkojix2, developer (self-described: Solo, first web app in Crystal), Personal blog post, DEV Community, published Mar 23, 2025
  108. What's new in @cline? (X post)Addy Osmani, engineer and author (self-described: Personal use of Cline in VS Code), X post, published Mar 5, 2025
  109. MCP MarketPlace Enterprise Usechezsmithy, for development teams in a regulated company (self-described: Highly regulated industry that vets every MCP server through infosec before release), GitHub issue, published Feb 20, 2025
  110. The Cline AI Assistant is MesmerizingMichael Lynch, professional developer (self-described: Solo, Zig command-line tool), Personal blog post, mtlynch.io, published Feb 1, 2025
  111. Why I use Cline for AI EngineeringAddy Osmani, engineer and author (self-described: Engineering work on complex systems), Newsletter, Elevate (addyo.substack.com), published Jan 30, 2025
  112. Using AI for Coding: My Journey with Cline and Large Language ModelsPaolo Galeone, independent developer (self-described: Solo, Go and Python SaaS side project), Personal blog post, P. Galeone's blog, published Jan 26, 2025

Jan 2025 to Oct 2026.

94 sources from practitioners, by month and sentiment.

The full readExplore all Cline Signals4 praise themes, 5 concerns, 5 points of agreement and 3 disagreements across all three groups, with all 112 sources open to inspect.

Where It Fits

Where does Cline belong in the stack?

Open Where It Fits: the Stack Field and the full map

Cline is Core across Build, supports Plan and Validate, and touches Ship and Operate.

Business area
Product & Engineering
Function
Software Development
Operating stack
Agentic Software Development
Primary role
Open-Source Coding Agent for the Editor and Terminal
Cline beside the rest of its stackThe Stack Field, every product on every layer
FootprintCline across the Agentic Software Development stack. Bar height is its role on each layer.

  1. Plan

    1. Work planning: Supports

      Work planning, supports. Plan mode reads the codebase and discusses an approach without changing files, and /deep-planning writes a detailed plan; planning stays in the team's tracker.

  2. Build

    1. Codebase context: Supports

      Codebase context, supports. Cline reads files, runs searches, and loads project rules on demand; it does not index or embed repository contents.

    2. Coding: Core

      Coding, core. This is Cline's primary stack role: editing code across a repository from the editor extension, the CLI, the desktop app, or a cloud agent session.

    3. Agent runtime: Core

      Agent runtime, core. The extension for VS Code and JetBrains IDEs is where developers approve, compare, and roll back agent steps; the CLI and desktop app run the same agent outside the editor.

    4. Tool execution: Core

      Tool execution, core. Cline runs terminal commands and MCP server tools behind per-category approvals, with an MCP Marketplace and Enterprise allowlists for which servers may run.

  3. Validate

    1. Debugging: Supports

      Debugging, supports. Cline can investigate failures by reading code and running commands, and its documentation recommends Plan mode for tricky bugs.

    2. Testing: Supports

      Testing, supports. Cline can write and run tests and fix failures, but it does not replace dedicated testing infrastructure.

    3. Code review: Touches

      Code review, touches. The Cline CLI can review pull requests from GitHub Actions, as Cline's documentation shows, but Cline is not a dedicated review product.

    4. Application security: Touches

      Application security, touches. Approval gating and admin controls limit what the agent can do, but Cline is not an application-security platform.

  4. Ship

    1. Source control: Touches

      Source control, touches. Checkpoints keep a shadow Git history for rollback and cloud agents open pull requests; Git stays the system of record.

    2. CI/CD: Touches

      CI/CD, touches. The Cline CLI runs headless in scripts and CI pipelines; delivery stays in the team's CI/CD.

  5. Operate

    1. Observability: Touches

      Observability, touches. Cline exports its own usage telemetry over OpenTelemetry and can help investigate production issues given context, but it does not provide observability for your systems.

Market mapThe layers Cline covers, and who else owns each one.

  1. Coding & implementation

    CoreBuild

  2. Agent interface & developer runtime

    CoreBuild

  3. Tool execution & integrations

    CoreBuild

  4. Work planning & task decomposition

    SupportsPlan

  5. Codebase context & intelligence

    SupportsBuild

7 more layers on the full map. Core, Supports and Touches describe coverage. They say nothing about quality.

Models

Which models run in Cline?

The current releases GILD tracks that public sources record running in Cline. Each model's page charts its public benchmark scores.

  1. GPT-5.4OpenAI's GPT-5.4 model for coding and professional work, with text and image input. API model gpt-5.4 (snapshot gpt-5.4-2026-03-05) at US$2.50 per million input tokens and US$15 per million output tokens, with a 1,050,000-token context window; prompts over 272K input tokens cost more. Also in Cursor; GitHub Copilot dropped it on October 19, 2026.GPT-5Benchmarks
  2. gpt-oss-120bOpenAI's largest open-weight model, 117B parameters with 5.1B active, with a 131,072-token context window. Weights under the Apache 2.0 license; OpenAI lists no API price for it, so it runs on your own hardware or a host's.gpt-ossBenchmarks
  3. MiMo-V2.6-ProXiaomi's flagship MiMo model, a mixture-of-experts model of about 1T parameters with 42B active and a 1M-token context. Weights under the MIT license.MiMoBenchmarks
  4. Qwen3.8 27BAlibaba's dense 27B Qwen3.8 vision-language model for coding and agentic work, with a 262,144-token native context. Weights under the Apache 2.0 license; on Alibaba Cloud Model Studio as qwen3.8-27b.QwenBenchmarks

More on Cline

  1. GILD ReviewGILD's evaluation, with the buying facts
  2. Research3 pieces
  3. Guides4 guides
  4. Operator Insights1 operator account
  5. Where It FitsOn the Agentic Software Development stack
  6. Related Products8 products