☑️Follow this @cursorcommunity account or @agidevice to stay in the loop about all things Cursor -
We are 23k+ Cursor a community members and growing!
The goal is to keep the Cursor community alive whether the communities on x shut down or not.
🚨Cursor 2.0 is here!
let's break it down ⬇️ - save this for later 📌
at a glance:
- Cursor launched Composer -the first coding model🔥
- a new (and beautiful) multi-agent interface!
- Embedded browser 🕸️
- Improved code review
and more!
Cursor agent can now work for longer! Plan mode!
How to use Plan Mode
Press Shift +
1. Answer clarifying questions on your requirements
2. Review or edit the detailed plan ( you can build directly right after)
3. Save the plan as a .md file in your repo
Cursor Updates - Have you used them?
1. add context like a git branch using the @ menu -
try @ branch
2. Custom slash commands -
- create reusable prompt
- Commands are stored in .cursor/commands/[command].md.
Run them by typing / in the Agent
Cursor 1.4 is out!
- More Agent steerability
- Improved Agent tools
- Separate models per Agent
- Usage and pricing visibility
- Compact Chat mode
- Sidebar for all agents
- Github support for Background agents
- Faster Background Agent start
More details below ⬇️-
AI forgetting during chat? @leerob breaks context down.
• Context stacks up fast
• Past context limit = blur
• System & user weigh the same
• Fresh chat → sharp focus
• Cursor flags overflow, trims smartly
I asked myself what does 10K tokens represent?
- Natural-language context:
• ~7500-8000 words (1 token ≈ ¾ word)
• ~15-20 single-spaced pages
• ~15 min of spoken dialogue at 120 wpm
- Code context:
• ~1000 lines (rule of thumb: 100 lines ≈ 1000 tokens)
🙌Breaking Amazing Cursor News🚨
@leerob joined @cursor_ai
this is a celebration and a win for all of us! 🎊 🎉
expect more education!
and an order magnitude of productivity will be unlocked for all of us with Lee’s support!