XNews

TRAE developer report and AI agent usage

TRAE developer report and AI agent usage

TRAE publishes an annual developer report with user metrics, usage patterns, and performance improvements. - Report lists total users: 6,000,000. - Report lists monthly active users: 1,600,000. - Weekly active users show near full weekday attendance. - TRAE iterated from Agent 1.0 to SOLO Beta to SOLO release. - Completion latency reduced by 60% per report. - Memory usage reduced by 43% per report. - Most-used tasks: code generation (~30%), bug fixes (~40%). - Less-used tasks: repo management, environment management, code optimization. - Top languages by perceived AI capability: Vue, Python, JavaScript, HTML, Java, TypeScript. - Platform supports over 11,000 MCP integrations. - Most-used AI programming mode: Builder (~80%); SOLO Coder (~40%). - Over 80% of users mix multiple agent modes. - A separate user asked whether Claude Code agents can automate monthly multi-source data collection into Excel (post 19314).

Key insights

TRAE report shows scale and strong user engagement

The TRAE developer report states a total user base of 6 million and monthly active users of 1.6 million. Weekly active usage is concentrated on weekdays, indicating consistent weekday engagement. TRAE has iterated rapidly from Agent 1.0 through SOLO Beta to the formal SOLO release, and the platform now supports over 11,000 MCP integrations. These metrics position TRAE as a widely used professional AI coding tool in its market.

Reported performance and usage breakdown highlights common tasks

According to the report, completion latency decreased by 60% and memory usage dropped by 43%. The most common user tasks are code generation (~30%) and bug fixing (~40%). Less-used features include repository management, environment management, and code optimization. The report also lists popular programming languages in user workflows ranked as Vue, Python, JavaScript, HTML, Java, and TypeScript.

User asks about automating multi-source data collection with agents

A user raised a practical question about automating monthly aggregation of exam data from multiple sources (Xiaohongshu, Tieba, Zhihu, school public accounts) into Excel, and asked whether an agent like Claude Code could perform the task. The question seeks a one-click or automated pipeline to collect and format distributed web/social content into structured spreadsheets.

Source posts