👋 How to start using Cline in 2 minutes (step-by-step thread 🧵)

So what do we think "Polaris Alpha" is?

NOW AVAILABLE! GLM 4.6 at 1000 token/s through the @cerebras provider on Cline. The best combination of speed and accuracy directly in your IDE or CLI.

Read more about hooks on our Docs: docs.cline.bot/features/hook…

Introducing Cline Hooks! Inject custom logic into your agentic workflows at key moments. Validate operations, monitor tool usage, and shape AI decisions automatically. What you can do: > Block problematic actions before they execute > Learn from operations and build project knowledge > Track everything for analytics or compliance > Trigger external tools at the right moments With six hooks covering the entire task lifecycle that go from starting a task all the way to post tool use.

Very excited to be included in The Information's 50 Most Promising Startups!

Live in Cline: kimi-k2-thinking

A good verifier is like a tea kettle. The only important question: Is your kettle whistling?

the tea kettle analogy goes so hard

Lost a hackathon in 2024. Won GitHub in 2025. With a 4,704% contributor growth Cline is the fastest growing AI-related open source project on GitHub.

call me @ npm install -g cline

.@MiniMax__AI M2 uses "interleaved thinking" - instead of one thinking block at the start, you get multiple thinking blocks throughout a single request. The model re-evaluates its approach as it goes, adapting based on tool outputs and new information. You'll see these thinking blocks appear in the UI showing its reasoning process in real-time. Available free in Cline until November 7!

Check out the changes we've made: github.com/cline/cline/blob/…

We've migrated our system prompt tool calling format to native tool calling and split that out for different model families. Here's why this results in a better experience using Cline: Models now return tool calls in their native JSON format, which they were specifically trained to produce. You'll notice fewer "invalid API response" errors - this particularly improves gpt-5-codex performance with significant reduction in failed operations. Parallel tool execution is now enabled. When Cline needs to read three files, it can execute them simultaneously instead of sequentially. System prompts are also smaller since tool definitions moved to API parameters, reducing token usage by approximately 15% per request. Native tool calling is currently supported for next-generation models: Claude 4+, Gemini 2.5, Grok 4, Grok Code, and GPT-5 (excluding gpt-5-chat) across supporting providers including Cline, Anthropic, Gemini, OpenRouter, xAI, OpenAI-native, and Vercel AI Gateway.

Changelog: github.com/cline/cline/blob/…

Full details: cline.bot/blog/cline-v3-35

And it's available for free in the Cline provider until November 7!

MiniMax M2 implements "interleaved thinking" - maintaining internal reasoning throughout task execution, not just at the beginning. You'll see multiple <thinking> blocks in the UI per API request showing the reasoning process in real-time.

When both "Read" and "Read (all)" are enabled, only "Read (all)" displays. Auto-approve is always enabled by default. We removed the toggle, favorites, and max requests.