AI Agent Hub
Back to plugins
dsh-codex-subscription preview

dsh-codex-subscription

Model Inference Updated 2026.09.15

Run the following command in DeepSeek Harness:

dsh plugin install imMamdouhaboammar/dsh-codex-subscription

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

Run dsh plugin install imMamdouhaboammar/dsh-codex-subscription in your DeepSeek Harness terminal to install the plugin; the source code is hosted at https://github.com/imMamdouhaboammar/dsh-codex-subscription

About this plugin

If you already pay for a ChatGPT Codex subscription but prefer DeepSeek Harness as your daily development environment, this plugin closes the gap cleanly. There is no OpenAI API key to paste and no separate Codex CLI to maintain. After a one-time browser sign-in inside DSH settings, Codex models, subscription search, live quota panels, and image generation all live directly in your DSH conversation window.

The core experience revolves around three pillars. First, transparent quota management: standard Codex, Spark, Credits, and other backend limits are listed separately with their reset windows and freshness timestamps. You can display usage as a percentage, a progress bar, or a Beta runway estimate, and every reset redemption requires a deliberate acknowledgement with a five-second cooldown, so a routing failure never silently falls back to a paid API. Second, creative tooling: a built-in image viewer supports zoom, pan, region annotation, and continue-editing hand-off to the composer; new and edited images return the exact original file path on the DSH host so an Agent can read or copy the artefact; and Fast mode is a one-click toggle in the model menu. Third, flexible context budgets: each model accepts a standard catalog window, an extended window, or a full numeric token limit, all of which feed DSH's native agent compaction policy without overwriting an unsaved draft.

This plugin is designed for developers who already hold a ChatGPT Codex subscription and want to keep every workflow inside DeepSeek Harness, whether that is interactive coding sessions, side-by-side multi-model comparisons, or one-shot headless tasks that print their answer and exit.

Screenshots

Use Cases

  • Invoke Codex models directly inside DSH conversations for coding and multi-step reasoning
  • Route search through the Codex subscription to leverage its web capabilities instead of the default engine
  • Annotate, zoom, and iteratively edit model-generated images within the built-in viewer

Best For

  • Developers who already hold a ChatGPT Codex subscription and want to use it inside DSH
  • Power users who need unified quota, context-window, and search routing management across models
  • Automation engineers running batch headless tasks inside DSH that print results and exit