AI Agent Hub
Back to plugins
⚙️

dayreel

Workflow Updated 2026.08.30

Run the following command in DeepSeek Harness:

dsh plugin install peikuo/dayreel

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install peikuo/dayreel in the DeepSeek Harness terminal to install the plugin; the source repository is available at https://github.com/peikuo/dayreel .

About this plugin

You spend a day juggling multiple sessions and workspaces inside dsh, and by evening the thread of what you actually did is scattered across transcripts. dayreel collapses that into a single sentence: type a prompt in the Web UI, and it pulls cross-session abstracts from the compaction layer, drills into individual sessions when detail is thin, then deterministically renders a styled HTML and Markdown daily report plus a narrated mp4 video. The entire stack is TypeScript and Node, with no Python dependency.

The plugin handles model routing on its own: at runtime it enumerates your LLM catalog, asks two quick questions (or reads a saved preference) to pick an authoring model, and falls back to a sensible default in headless mode, so adding or removing models never requires a plugin config change. When the video is ready, an inline player with a seekable progress bar appears right in the tool-call row, no URL copying needed. Verbal feedback you give afterward is passively logged and feeds into a force regeneration on the next run, iteratively sharpening the recap.

It is built for developers who run multiple projects, models, and sessions in dsh each day and would rather watch a thirty-second narrated recap and read a well-formatted report than scroll through chat history.

Use Cases

  • Quickly recap a day of work across multiple projects in dsh
  • Replace scrolling through chat logs with a narrated summary video
  • Archive and visualize cross-workspace session activity at a glance

Best For

  • Developers juggling multiple projects and sessions daily in dsh
  • Engineers who prefer a video recap over reviewing raw text logs
  • Teams wanting a lightweight daily report tool with a pure TypeScript stack