TRAE

What is TRAE?

TRAE, an AI IDE programming tool launched by ByteDance , is a highly intelligent AI development engineer. Integrating an AI Coding Agent, it can understand requirements, utilize tools, and independently complete various development tasks. Through the TRAE SOLO mode, TRAE understands developer needs and automates the entire process from code writing to testing, significantly improving development efficiency. TRAE supports natural language dialogue, allowing developers to interact with the tool using everyday language, simplifying complex tasks. TRAE prioritizes user privacy, employing a local-first strategy and a principle of minimal data collection to ensure data security. TRAE enhances productivity, allowing developers to focus more on innovation and problem-solving, freeing them from tedious coding tasks.

TRAE’s main functions

• AI Coding Agent : An integrated AI assistant that understands development needs, provides intelligent question answering, code auto-completion, and agent-based automatic programming capabilities.
• TRAE SOLO mode : A highly automated development approach where AI drives the entire process from requirements understanding, code generation, testing to result preview.
• Intelligent IDE : Provides traditional IDE functions such as code writing, project management, plugin management, and source code management, while integrating AI into the workflow.
• Customizable AI Agents : Users can flexibly define tools, skills, and task logic according to their needs to build their own AI agent team.
• Tool customization : Supports seamless integration of various external tools, expanding the capabilities of intelligent agents.
• More contextual understanding : The tool can understand code repositories and terminal information, supports online searches, uploads of document sets, and configuration of AI working rules.
• Cue intelligent prediction : Based on an optimized model, it achieves accurate continuation writing, intelligently perceives the repository context and editing trajectory, and automatically predicts the next change point.
• Privacy and Security : We adhere to the principles of “local priority” and “minimize data collection” to protect user privacy and data security.
• Regionalized deployment : User data and service infrastructure are stored and deployed based on the account’s location to meet local data regulations.

How to use TRAE

• Download and install : Visit the TRAE official website (international version: trae.ai), download the client, and install it on your operating system (supports macOS 10.15 and above, Windows 10 and 11).
• Startup and Login : Start TRAE IDE, register an account and log in.
• Create or import projects : Create or import projects by importing local folders, cloning repositories from GitHub, or cloning directly from a Git repository URL.
• Choose your work mode : Based on your development needs, choose either IDE mode or SOLO mode. SOLO mode allows AI to lead the development process, while IDE mode retains more control.
• Use the AI ​​assistant : Interact with the AI ​​assistant in side or inline conversations to get coding help, code explanations, bug fixes, and more.
• Manage plugins : Install, disable, or uninstall plugins as needed to extend TRAE’s functionality.
• Privacy settings : Enable privacy mode if needed to prevent TRAE from storing or using chat interaction content.
• Code writing and editing : Use TRAE’s intelligent programming tools, such as CUE, to assist with code writing and editing.
• Handling code changes : Accept or reject code changes generated by the AI ​​assistant, and manage code modifications in the project.
• Model management : Switch between different AI models as needed, or add custom models to meet specific requirements.
• Explore more features : Use TRAE’s intelligent agents, contextual understanding, and other features to further improve development efficiency.

TRAE’s product pricing

• TRAE (domestic version)
  • TRAE Personal Edition : Individual users can use all the features of the domestic version without paying, including integrated domestic AI models (such as Doubao, DeepSeek, etc.) and basic functions such as code generation and intelligent completion.
  • TRAE Enterprise Edition :
    • Seat fee : 69 RMB per seat per month. Benefits include 500 Q&A sessions (Agent/Chat/Builder) per month. Unlimited uses for autocomplete, InlineChat, Fast Apply, and AI Fix.
    • Pay-as-you-go billing : When a member’s resource usage exceeds the package limit, pay-as-you-go billing must be activated. Billing is processed hourly, and there may be an hourly delay in billing.
      • Billing formula : Bean Bun Model Service Fee = Input Token Usage × Input Token Unit Price + Output Token Usage × Output Token Unit Price.
• TRAE International Version
  • Free : The free version offers 10 fast requests and 50 slow advanced model requests, 1000 advanced model requests, and unlimited autocomplete per month.
    Pro : Professional version, $3 for the first month, $10 per month thereafter. Offers 600 fast requests and unlimited slow premium model requests, 300 extra fast requests (limited time offer), unlimited premium model requests, and unlimited autocomplete.

Application scenarios of TRAE

• Code writing and editing : Use TRAE’s AI assistant for automatic code completion and error correction to improve coding efficiency and accuracy.
• Project Management : Use TRAE to manage project files, version control, and task assignment, simplifying the project management process.
• Intelligent Q&A and Learning : While writing code, you can get code explanations, comments, and bug fixing suggestions by talking to the AI ​​assistant, accelerating the learning process.
• Code review and optimization : Use TRAE’s intelligent analysis tools to review code, identify potential problems, and propose optimization suggestions.
• Automated testing : Use TRAE to automatically generate test cases, execute tests, analyze results, and improve testing efficiency.