Trae is an AI-native, VS Code-based IDE built by ByteDance, offering free access to a generous tier of AI coding features alongside paid plans for heavier use. It officially launched on January 19, 2025, with an international English/Chinese version following on January 20, and a China-specific version — running on ByteDance's own Doubao models plus DeepSeek — launched separately on March 3, 2025.
Who builds it
- Built by ByteDance, the Chinese technology company also known for TikTok and Doubao.
- Positioned under the tagline "Collaborate with Intelligence," Trae ships two related products: TRAE IDE for coding and TRAE Work, a broader AI work assistant.
- ByteDance's VP of Technology, Hong Dingkun, stated Trae's monthly active users had exceeded 1 million as of June 12, 2025.
Core features
- Built on the VS Code foundation, so existing extensions and workflows carry over.
- Builder mode for AI-assisted project scaffolding from natural-language instructions.
- SOLO mode, a standalone app experience included on every plan, from Free through Ultra.
- Inline autocompletion, with monthly request allowances that scale by plan.
- Concurrent cloud agent tasks, also scaling with plan tier, plus queue-priority differences between plans.
Pricing
- Free: 5,000 autocompletion requests per month, 2 concurrent cloud tasks, standard queue priority, SOLO mode included.
- Lite: $3/month, unlimited autocompletion, 2 concurrent cloud tasks, fast queue priority.
- Pro: $10/month (7-day free trial), unlimited autocompletion, 10 concurrent cloud tasks, fast queue priority, limited early access to new models.
- Pro+: $30/month, roughly 3.5x the usage allowance of Pro, 15 concurrent cloud tasks, fast queue priority.
- Ultra: $100/month, roughly 20x the usage allowance of Pro, 20 concurrent cloud tasks, model early access.
Who it's for
Trae's generous free tier and low-cost entry plans make it attractive to individual developers, students, and anyone prototyping quickly who wants a familiar VS Code-style editor with AI built in at no or low cost. Because it is built and operated by ByteDance, teams with strict requirements around data residency or vendor jurisdiction should factor that ownership into their evaluation alongside the feature set, in the same way they would for any AI IDE run by a large platform company.