DeepSeek TUI's project-owned surfaces have entered a CodeWhale rename wave, while the public site still teaches legacy `deepseek` commands

Accepted developer update

Today's safe publishable change belongs on the DeepSeek TUI track again, but this time the important delta is not just another patch release. The stronger upstream signal is a naming split across project-owned surfaces that directly affects how developers copy install commands.

What changed upstream

• The public website still presents the legacy path: the DeepSeek TUI homepage and install page still show the DeepSeek TUI brand, the deepseek command, the deepseek-tui npm package, and v0.8.40 as the latest stable release.
• The live project activity feed now shows a CodeWhale rename sequence: the same official website currently highlights recent upstream work titled roughly as README/docs rename to CodeWhale, npm publish as codewhale while keeping deepseek-tui as a deprecation shim, and binary renames that keep deepseek aliases.
• GitHub has already moved further than the public site: the canonical GitHub repo path for the project now redirects from Hmbown/DeepSeek-TUI to Hmbown/CodeWhale.
• Compatibility remains the explicit bridge: the rename wave is not described as an overnight hard cut. The project-owned surfaces indicate compatibility aliases are being kept so existing deepseek users do not break immediately.

Why this matters for crawlable setup pages

This is a better developer signal than a vanity version bump because it changes what a technically trustworthy setup page should tell readers today:

1. Do not pretend there is only one canonical command spelling across all official-looking pages right now.
2. If you follow the public website today, the documented install path is still deepseek / deepseek-tui.
3. If you follow the redirected GitHub repo and newer rename work, you will start seeing codewhale naming.
4. Safer setup advice should tell readers to stay on one verified command surface at a time and not mix package names, binary names, and repo paths from different rename stages.

What we rejected today

• DeepSeek local development: the official DeepSeek V4 Flash Hugging Face card still exposes the same maintained baseline already covered on this site: vLLM and SGLang launch commands, Docker Model Runner, quantization browsing for llama.cpp, Ollama, and LM Studio, and local curl validation. No stronger official local-runtime delta today.
• DeepSeek in Claude Code / Cloud Code: DeepSeek's official Claude Code guide still documents the same Anthropic-compatible environment recipe already reflected in the maintained guide, including deepseek-v4-pro[1m], deepseek-v4-flash, CLAUDE_CODE_SUBAGENT_MODEL, and CLAUDE_CODE_EFFORT_LEVEL=max.
• Community chatter about pricing or cache-hit behavior: useful discovery only, but not strong enough to publish without a clearer official docs change tied to one of today's three target tracks.

Editorial takeaway

Today's public update belongs on the DeepSeek TUI guide, the dedicated /deepseek-tui landing page, and the news stream. It does not affect stocked plans, pricing cards, stock, or any purchasable inventory surface.

Sources checked

• DeepSeek TUI homepage
• DeepSeek TUI install page
• CodeWhale GitHub repository
• DeepSeek official Claude Code integration guide
• DeepSeek V4 Flash model card