AI Agent Hub
Back to plugins
dsh-plugin-oauth-subs preview

dsh-plugin-oauth-subs

Model Inference Updated 2026.09.05

Run the following command in DeepSeek Harness:

dsh plugin install xxww0098/dsh-plugin-oauth-subs

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

Run dsh plugin install xxww0098/dsh-plugin-oauth-subs in your terminal to install; the source is at https://github.com/xxww0098/dsh-plugin-oauth-subs

About this plugin

DeepSeek Harness gives you a unified inference workflow, but your paid AI subscriptions are usually scattered across a dozen or more platforms: ChatGPT Codex, xAI Grok, Kimi Code, GitHub Copilot, Cursor, Ollama Cloud, GLM, Kiro, Antigravity, and others. dsh-plugin-oauth-subs bridges that gap by plugging your existing OAuth sessions or API keys directly into Harness, so you no longer need a separate API allotment for every model you already pay for.

The plugin auto-detects the right auth flow per provider: PKCE for Codex, device-code for Grok and Copilot, local credential import for Kimi and Cursor, ZCode identity for GLM, and plain API keys for Ollama Cloud or Kiro. After login it spins up a loopback proxy (default port 8318) that routes llm-pi-ai calls to the correct upstream protocol—openai-responses for Codex and Grok, anthropic-messages for GLM, openai-completions for the rest. Your Harness sessions, tool pipelines, and compaction logic stay exactly as they are; only the underlying tokens now come from your existing subscriptions. The settings panel lets you toggle models per family, set reasoning depth from low to max, and enable Fast or 900K long-context modes on demand.

If you already hold two or more AI subscriptions, the most tangible win is eliminating redundant spend while keeping the full Harness workflow intact. Every subscription card also shows live quota usage and reset times at a glance, so you can catch rate-limit pressure before it derails a critical task.

Screenshots

Use Cases

  • Use existing ChatGPT, Grok, or Kimi subscriptions directly inside Harness
  • Manage multi-provider AI auth, model switching, and quota in one place
  • Cut redundant API costs by reusing already-paid subscriptions

Best For

  • Developers holding two or more AI subscriptions simultaneously
  • Teams that want a unified workflow while preserving per-platform quotas
  • Power users needing fine-grained control over models, reasoning depth, and context window