Cursor
Cursor is Anysphere's AI IDE, a VS Code fork with tab complete, inline edit, and agents. Rechecked 2026-08-17: cursor.com/pricing sells Hobby $0, Individual $20 / mo, Pro $40 / mo, Teams $40 / user / mo. Cloud Agents and Bugbot are listed as add-ons, not as the seat. This page used to claim 17.48M visits, 1M users, $500M ARR, and a 40%+ speedup. Those figures are not on the live pricing page, so they are out.
Compare GitHub Copilot if you want the GitHub-native suite, and Windsurf if you are looking at Cognition's IDE instead.
Key Features
- IDE, not a plugin: A standalone editor with Tab, inline edit, and agent modes.
- Two agent surfaces: The pricing page distinguishes Agents in the editor from Cloud Agents.
- Bugbot add-on: On-demand $40 / mo, or $40 per 100 PRs. Teams include a $32 / user / mo credit toward that.
- Rules / project context: Still the way teams steer the model. Recheck the current rules file name in-app; do not freeze
.cursorrulesas the only path. - Model picker: Live model list moves. Do not treat Claude Opus 4 / GPT-4.1 / Gemini 2.5 Pro as the 2026 catalog.
Use Cases
- Daily IDE work where tab complete and an in-editor agent matter more than a browser chat.
- Teams that will pay $40 / user and maybe Bugbot on PRs.
- People comparing Copilot vs a dedicated AI IDE.
Limitation: Hobby is not a full Pro. Individual vs Pro is a usage-limit split, not two different products. We did not re-verify Auto / Composer Premium token rates beyond the on-demand table.
Pricing
cursor.com/pricing on 2026-08-17:
| Plan | Price | Notes from that page |
|---|---|---|
| Hobby | $0 | Limited Agent requests and tab complete. No credit card. |
| Individual | $20 / mo | Unlimited tab. Agent limits. |
| Pro | $40 / mo | Higher Agent usage. $20 / mo of included Auto + Composer Premium tokens, then $1.25 / 1M input and $6.00 / 1M output on demand. |
| Teams | $40 / user / mo | Shared chats, centralized billing, usage analytics, SSO. Roles on Enterprise. |
| Cloud Agents | Usage | $0.25 / 1M cached input, $1.25 / 1M input, $6.00 / 1M output. |
| Bugbot | $40 / mo on-demand or $40 / 100 PRs | Teams include $32 / user / mo Bugbot credit. |
Getting Started
- Download Cursor from cursor.com.
- Start on Hobby only to see Tab. Pay Individual or Pro before you lean on Agent.
- Add project rules before you let the agent touch more than one file.
- Recheck Cloud Agents and Bugbot as separate bills.
Frequently Asked Questions
Is Pro still $20?
No. Individual is $20. Pro is $40 on the page we opened.
Are Cloud Agents included?
They are listed as a usage add-on.
Did Cursor buy Continue?
This catalog already covers Continue separately. Do not mix the two products here.
Alternatives
- GitHub Copilot: Completions and a cloud agent inside GitHub.
- Windsurf: Cognition IDE if you want that stack.
- Claude Code: Terminal-first agent if you do not want another IDE.
Tips
- Quote $20 / $40 from the pricing page, not old $20-Pro tweets.
- Treat Bugbot and Cloud Agents as extra meters.
- Recheck model names in the picker. Launch-era Opus 4 copy is stale.
Conclusion
Cursor is still the paid AI IDE, now with Individual $20, Pro $40, and add-on Cloud Agents / Bugbot. Start at cursor.com/pricing, then compare Copilot if you already live in GitHub.
Comments
No comments yet. Be the first to comment!
Related Tools
aiXcoder
www.aixcoder.com
Chinese AI coding assistant with an Apache-2.0 7B model. No public SaaS price table; treat cloud seats as sales-quoted.
Bolt.new
bolt.new
StackBlitz in-browser app builder. Rechecked: Free $0 (1M tokens/month), Pro $25, Teams $30. Not an unlimited no-install slogan.
Origin (by Cursor)
cursor.com/docs/origin
Cursor's Git forge and code-hosting platform built for AI coding agents, with repos, pull requests, GitHub sync, and agents in every repo.
Related Insights
Cursor vs GitHub Copilot: Complete Comparison 2026
In-depth comparison of Cursor and GitHub Copilot. Discover which AI coding assistant is best for your workflow with detailed feature analysis, pricing, and real-world testing.
Seven AI Coding CLIs, Six Months: No Matter How Strong the Model, Work Needs Supervision
Claude Code, Codex, opencode, pi, omp and DeepSeek Harness all have personalities. After six months of deep use I run a division of labor: pi for the fastest cheapest reviews, omp for complex PRs, DeepSeek Harness on V4 Flash for high-frequency low-cost review, and Claude Code, Qoder and Cursor for writing. No matter how strong the model, work needs supervision — ideally from an independent third party.
Skills + Hooks + Plugins: How Anthropic Redefined AI Coding Tool Extensibility
An in-depth analysis of Claude Code's trinity architecture of Skills, Hooks, and Plugins. Explore why this design is more advanced than GitHub Copilot and Cursor, and how it redefines AI coding tool extensibility through open standards.