AI Agent Hub
Back to plugins
🖥️

dsh-conversation-nav

Client Updated 2026.08.24

Run the following command in DeepSeek Harness:

dsh plugin install Jaeger0624/dsh-conversation-nav

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

Run dsh plugin install Jaeger0624/dsh-conversation-nav in the DeepSeek Harness terminal to install; full source at https://github.com/Jaeger0624/dsh-conversation-nav .

About this plugin

Once a conversation in the DeepSeek Harness Web GUI stretches past a few dozen turns, scrolling back to a specific message becomes a tedious, memory-dependent exercise. dsh-conversation-nav brings the signature piano-key rail from the Codex desktop client into the browser: a vertical column of ticks along the left edge of the chat area, one per user message, turning any long thread into an instantly navigable timeline.

Hover over a tick to reveal a detail card with a one-line message summary, up to three lines of the model reply preview, and a meta line showing send time (HH:mm for today, M-dd HH:mm this year, full date across years), round duration, and output token count. Click to smooth-scroll to that turn with a brief highlight; while scrolling, the tick for the turn you are reading darkens in real time so you always know your position. Right-click a tick to cycle through amber, red, blue, and green markers, optionally attach a text note, and all marks persist per session in localStorage across reloads. The plugin is fully event-driven with no persistent timers, so even very long conversations stay fluid.

If you routinely work through dozens or hundreds of turns with DeepSeek in the web interface and need to revisit, compare, or annotate key exchanges, this zero-dependency, purely client-side plugin gives you that navigation out of the box.

Use Cases

  • Jumping to a specific turn in a long thread without endless scrolling
  • Reviewing round duration, token count, and exact timestamp of past messages
  • Color-coding and annotating key turns for later reflection

Best For

  • Developers who regularly run multi-turn technical sessions
  • Teams or individuals reviewing AI conversation quality and efficiency
  • Web users who prefer the Codex desktop interaction style