正在加载视频...

视频加载失败

Introducing Cline CLI 2.0: An open-source AI coding agent that runs entirely in your terminal. Parallel agents, headless CI/CD pipelines, ACP support for any editor, and a completely redesigned developer experience. Minimax M2.5 and Kimi K2.5 are free to use for a limited time. From prompt to production. All...

552,567 次观看 • 8 个月前 •via X (Twitter)

64 条评论

Cline 的头像
Cline8 个月前

We built CLI 2.0 by listening. The community told us the terminal needed to be a first-class surface for AI coding, not just an afterthought. We studied what other tools were doing, looked at how developers actually work in their terminals, and redesigned the entire experience around that.

Cline 的头像
Cline8 个月前

What's new in Cline CLI 2.0: + Completely redesigned terminal UI with interactive mode + Parallel agents with isolated state per instance without manual instance creation. + Improved headless mode for CI/CD pipelines + Added ACP support for Zed, Neovim, and Emacs

Cline 的头像
Cline8 个月前

`npm install -g cline` and start building. Now available on Windows, Mac, and Linux. Read more:

NewAIWorld 的头像
NewAIWorld8 个月前

Parallel agents mean I can now produce four times as many bugs in the same amount of time

Cline 的头像
Cline8 个月前

no one can stop you now

Minh-Phuc Tran 的头像
Minh-Phuc Tran8 个月前

You got me with the tmux support! I didn't know that TUI can follow your mouse like that either, so dang cool 😄

Cline 的头像
Cline8 个月前

honestly best feature of the CLI

Raul Junco 的头像
Raul Junco8 个月前

This was missing: open-source, model-agnostic, editor-agnostic. 👌

Cline 的头像
Cline8 个月前

don't get locked-in to a single provider or model

Raul Junco 的头像
Raul Junco8 个月前

This is the way!

🚨 AI News | TestingCatalog 的头像
🚨 AI News | TestingCatalog8 个月前

👀👀👀

Cline 的头像
Cline8 个月前

👀

darren 的头像
darren8 个月前

coming to @tembo asap

Cline 的头像
Cline8 个月前

@tembo delegate your coding tasks to us

Rohan Paul 的头像
Rohan Paul8 个月前

The AI dev space is really on fire recently. Love the simple installation with `npm install -g cline`

Cline 的头像
Cline8 个月前

it really is that simple 🙌

Marcus Webb 的头像
Marcus Webb8 个月前

Headless mode for CI/CD is the killer feature here. Terminal-first dev tools are long overdue for proper AI integration.

Tech with Mak 的头像
Tech with Mak8 个月前

Terminal is where real work happens. Great stuff!

Cline 的头像
Cline8 个月前

the original primitive

Kris 👾 的头像
Kris 👾8 个月前

Hey quick question, when using those free models for a limited time do they collect the data? I mean do they store input and outputs for future training?

Cline 的头像
Cline8 个月前

We don't store data ourselves but model providers have different ToS that might treat this differently

Stein 的头像
Stein8 个月前

@BanterBytes @grok which models and what data collects the free llms used in cline?

Grok 的头像
Grok8 个月前

The free LLMs in Cline CLI 2.0 are Minimax M2.5 and Kimi K2.5. Minimax collects inputs/outputs, account info, and usage data; anonymized data may be used for model training. Opt-out via account deletion or consent withdrawal. Kimi collects inputs/outputs, device/usage data; used for service improvement including model training. Opt-out by deleting account or contacting support. Check providers' ToS for details, as Cline doesn't store data itself.

Stein 的头像
Stein8 个月前

@cline @BanterBytes So it means it can collect private data? Like apis and stuff?

Grok 的头像
Grok8 个月前

Yes, if you input private data like API keys into prompts, it could be collected as part of inputs/outputs. Both Minimax and Kimi's policies allow using such data for training/improvements unless you opt out. Avoid sharing sensitive info; use local models for privacy. Check their ToS for specifics.

Stein 的头像
Stein8 个月前

@cline @BanterBytes Are Kimi and mini found for local?

Grok 的头像
Grok8 个月前

Yes, both Kimi K2.5 and Minimax M2.5 are available for local running. Minimax M2.5: Download from Hugging Face (MiniMaxAI/MiniMax-M2.5) and run with tools like llama.cpp or vLLM. Requires significant RAM/VRAM (e.g., 100GB+ for quantized versions). Kimi K2.5: Quantized versions on Hugging Face (unsloth/Kimi-K2.5-GGUF); run via llama.cpp. Needs 240GB+ unified memory for best performance. Check hardware reqs and guides on Unsloth or HF.

Ritika Agrawal 的头像
Ritika Agrawal8 个月前

The terminal UI looks fantastic and the agents having isolated state per instance will speed up my work exponentially! I will definitely check this out.

Etisha Garg 的头像
Etisha Garg8 个月前

ACP support feature looks really interesting to me

Cline 的头像
Cline8 个月前

bring us to your favorite IDE 🫶

Ankur💻🎧💪 的头像
Ankur💻🎧💪8 个月前

congrats team.

Cline 的头像
Cline8 个月前

🫡

Thalia 🌸 的头像
Thalia 🌸8 个月前

parallel agents in the terminal is a game changer. the future isn't one AI doing everything — it's specialized agents coordinating. coding agents, business agents, research agents, all talking to each other. excited to see where ACP interop goes.

Mayank Gupta🧑🏻‍💻 的头像
Mayank Gupta🧑🏻‍💻8 个月前

the video makes me play around with the CLI now, will checkout tonight

Cline 的头像
Cline8 个月前

let us know what you build!

Carlos Troyana 的头像
Carlos Troyana7 个月前

Parallel agents in CI/CD pipelines is powerful, and it raises the question no one is asking yet: how does each agent verify the origin of its instructions? Built @meshsig to answer that. Ed25519 signatures + did:msig: identity for every agent. Works natively with Cline via MCP.

Marko Denic 的头像
Marko Denic8 个月前

Sounds interesting. Thanks for sharing.

AshutoshShrivastava 的头像
AshutoshShrivastava8 个月前

Never a dull day in the AI space.. 👏

Cline 的头像
Cline8 个月前

🫡

Pratham 的头像
Pratham8 个月前

I know what I need to play around with over the weekend.

Cline 的头像
Cline8 个月前

let us know what you build

Petar Ivanov 的头像
Petar Ivanov8 个月前

That's a massive upgrade! Congrats to the whole team!

joelhan 的头像
joelhan8 个月前

I love the UI design. Looks dope

Cline 的头像
Cline8 个月前

thx 🫶

Isaac Tony 的头像
Isaac Tony8 个月前

Good stuff, cannot wait to try the parallel agents feature 🚀

Somi 的头像
Somi8 个月前

oh parallel agents in the terminal is exactly what I've been wanting. been running everything through a single agent and the context switching overhead was killing me

Jeet 的头像
Jeet8 个月前

looks interesting, gotta try this out!

Cline 的头像
Cline8 个月前

let us know how it goes! All feedback is welcomed

Leviath 的头像
Leviath8 个月前

why is kimi 2.5 charging me if its free...

Cline 的头像
Cline8 个月前

it should show $0 in your account dashboard.

exitLQ 的头像
exitLQ8 个月前

@Scobleizer and another one to install and try lfg

Abdulmuiz Adeyemo 的头像
Abdulmuiz Adeyemo8 个月前

Wow Terminal is really great to coding

Jay Sensei👾 的头像
Jay Sensei👾8 个月前

Been using @cline for long time and I must say they are community driven. Congrats guys 👍🏻 @warpdotdev your next move 👀

Ungaanalysis 的头像
Ungaanalysis8 个月前

It's super clean! need that bug mode...

Jaynit Makwana 的头像
Jaynit Makwana8 个月前

Cline CLI 2.0 is insane🔥

Cline 的头像
Cline8 个月前

we've been cooking lately

Squiggle 的头像
Squiggle8 个月前

wow, cline cli 2.0 sounds huge! can it write tests too? 🔥

Cline 的头像
Cline8 个月前

🔥

Markandey Sharma 的头像
Markandey Sharma8 个月前

This is huge for terminal-first developers. Running an AI coding agent fully in CLI feels like the right direction.

Faizal 的头像
Faizal8 个月前

This looks super handy, especially the parallel agents and CI/CD setup in the terminal. Is there a quickstart for getting it running in GitHub Actions

Cline 的头像
Cline8 个月前

cooking on this 🍲

Amira Zairi 的头像
Amira Zairi8 个月前

Congrats 🎉

Sumanth 的头像
Sumanth8 个月前

Good to see ACP support. Can't wait to give it a try!

Cline 的头像
Cline8 个月前

🫶 keep us posted

相关视频