TRAE Launches Developer Personal Assistant TRAE Bao, Introducing a Dual-Layer Global and Project Assistant

Top banner ad space

You are using an old version of IE browser. For the security of your Leiphone account and a better product experience, we strongly recommend using a faster and safer browser.

Lead-in: On October 29, the AI development product TRAE officially launched the developer personal assistant "TRAE Bao," introducing a dual-layer assistant system of "global assistant + project assistant" into the development workflow.

On October 29, the AI development product TRAE officially launched the developer personal assistant "TRAE Bao." Unlike the previous Agent conversations centered on a single task, TRAE Bao introduces a dual-layer assistant system into the development workflow, enabling AI to continuously understand the projects, tasks, and decisions that developers accumulate over time, rather than merely responding to a single current instruction.

For a long time, the time users spend in the AI development process has been consumed not only by writing code itself, but also in large part by project breakdown, task assignment, and switching back and forth between multiple sessions. When a new task starts, developers need to brief the AI again on the project background, tech stack, and historical decisions; once requirements enter a complex stage, they must break down tasks themselves, switch sessions, track progress, and then assemble the results into a complete solution.

TRAE Bao adopts a dual-layer architecture of "global assistant + project assistant," corresponding to the above two types of scenarios respectively. The "global assistant" handles cross-project memory and coordination; it can read developers' past tasks, conversations, and outputs, and proactively sort out the reusable solutions and constraints within them. When starting a new task, developers only need to state the goal, and the assistant can locate relevant experience from historical projects and complete the automatic positioning of context. The "project assistant" focuses on the parallel execution of a single project. Based on the precise context of the current project, it breaks complex requirements into multiple executable tasks, schedules subtasks to advance in parallel, integrate, and summarize, while developers complete control by "listening to reports and making judgments."

Specifically, by opening the TRAE desktop client, you can enter the global assistant through "TRAE Bao," and by creating or opening a project you can enter the project-level assistant conversation. The global assistant supports binding to instant messaging tools such as Feishu and WeChat, allowing tasks to be dispatched and progress to be synced in real time without opening the client. The current version supports local tasks; cloud tasks and mobile capabilities will be opened in subsequent iterations.

AsiaInfo × Volcano Engine: 6,000+ seats, using TRAE to run enterprise-level AI research ...

NIO × Volcano Engine: TRAE goes deep into cockpit R&D, from personal efficiency ...

Content creators don't know how to build AI workflows? TRAE Work made this 400,000 ...

JD procurement and sales employee sued for publishing rumors to defame other platforms; the court has filed the case

From Dreamore to CreaXene: AI generation tools are moving toward creative workflows

Doubao Work adds a canvas feature and connects to the Doubao 2.1 Lite model

Latest articles

Alibaba brings Double 11 to South Korea with AliExpress, and even gives away 4 Teslas!

Shunheng Intelligence's self-developed dexterous humanoid robot SynDexy C1 achieves small-batch delivery for the first time

Douyin's Li Liang posted: Hope peers compete in the sunshine, and expert self-media should abide by basic moral norms

Blockchain Nvidia Social networks Silicon Valley Smart speakers Data security Routers Google Glass xbox Kindle Fire Galaxy S4

Set now Later

Youlu tracking code