Tokenwork.app, a GPT API platform, has launched its fourth public beta, claiming support for GPT-5.5, GPT-5.4, and Codex. New users receive a $1 credit after completing WeChat verification, and referrers earn $1 plus a 5% commission on the invited user's top-ups. Users who comment on the promotional post with a base64-encoded email are offered an extra $5 credit. The operator states the platform is built primarily for their own long-term use, with testing focused on stability, Codex compatibility, API response speed, peak-hour availability, and billing accuracy rather than low pricing.
A v2ex user listed multiple closed-source coding agent products (Tencent's Qclaw Workbuddy, ByteDance's Trae Work and 扣子编程, OpenAI's Codex, Anthropic's Claude) and asked whether open-source alternatives exist. They want to study how such agents internally call MCP skills and encapsulate Python and Node.js environments. The post is a request for learning resources and discussion.
A developer on V2EX posted five practical tips for using Claude in software development, including establishing a CLAUDE.md file as a project constitution that sets thinking and action rules, maintaining a separate file for project pitfalls, cross-document auditing with Claude, using Codex for adversarial code reviews, and designing frontend systems via claude.ai/design for style consistency. The post includes the full content of the user's CLAUDE.md, which contains nine concrete principles organized into Thinking, Action, Audit, and Communication sections, emphasizing first-principles reasoning, extreme quality, reading before modifying, mandatory self-audits, and direct communication without flattery. The CLAUDE.md explicitly takes precedence over any conversational instructions.
A V2EX user argues that the initial peak of AI coding hype has passed, based on four personal observations. Reported resale prices for ChatGPT Plus and Gemini Pro subscriptions have collapsed to around ¥15 and ¥12 respectively, signaling reduced demand. The share of AI-related topics on the V2EX forum fell from nearly 100% to a lower level. The user's own token consumption decreased after initial project stabilization. Additionally, interest in 'vibe coding' on Xiaohongshu has waned as non-coders who quickly built apps no longer generate new ideas to sustain activity.
VibeTrail is a macOS open-source utility (Rust + Tauri) that reads local session files from Claude Code, Codex, and Antigravity to provide a unified interface for managing AI coding agent conversations. It aggregates projects by working directory, offers full-text search across all agent sessions using ripgrep, and enables one-click resume in terminal emulators like iTerm2 and Warp. The tool is read-only, requires no database or background process, and opens sessions with 20,000 records in 0.06 seconds. It is released under Apache 2.0 with a signed and notarized dmg available on GitHub.
A user on V2EX posted a promotional thread inviting community members to share their experiences and tips regarding account bans on AI models such as Claude, GPT, Gemini, and DeepSeek. The thread serves as a discussion hub for users facing bans on these large language model services, under the name '大模型渡劫' (Large Model Tribulation). The thread's content is purely a call for discussion and does not contain any specific news, findings, or product announcements.