What it shows

Highlights

• Surface-aware Discord profile switching for Codex CLI / VS Code and Codex App.
• Fast mode visibility driven by Codex global state, with a lightning-marked model label.
• Reasoning effort visibility from live turn_context telemetry, including Extra High.
• Full-screen plan selector in the TUI, with instant save to config.
• Stable activity interpretation for long sessions, commentary turns, waiting states, and idle transitions.
• Deterministic active-session ranking when multiple Codex sessions are open at once.
• Compact but information-dense Discord payload formatting that stays within Discord limits.
• Local build and release layout that keeps Cargo cache under .build/target and final artifacts under releases/<platform>/.

Quick Start

1. Configure Discord assets

Create or update your Discord Developer applications with the image keys used by the runtime:

• Codex CLI / VS Code profile:
  • codex-logo
  • openai
• Codex App desktop profile:
  • codex-app
  • openai

2. Build

• Windows:

./scripts/build-release.ps1

• Linux / macOS:

./scripts/build-release.sh

Or build directly with Cargo:

cargo build --release

3. Run

codex-discord-presence

4. Use the TUI

While the terminal UI is open:

• Press P to open the full-screen account plan selector.
• Use arrow keys or 1-7 to select:
  • Auto Detect
  • Free
  • Go
  • Plus
  • Pro
  • Business
  • Enterprise
• Press Enter to apply and save immediately.
• Press P or Esc to close without changing the current plan.
• Press q or Ctrl+C to quit the runtime.

5. Validate health

codex-discord-presence status
codex-discord-presence doctor

Feature Spotlight

Fast mode visibility

The runtime reads ~/.codex/.codex-global-state.json and checks:

• electron-persisted-atom-state.default-service-tier

If the value is fast, the model is shown with a lightning prefix:

⚡ GPT-5.5

Reasoning effort visibility

Reasoning effort is resolved from the active session in this order:

1. turn_context.payload.effort
2. turn_context.payload.collaboration_mode.settings.reasoning_effort

Rendered examples:

GPT-5.5 (Low)
GPT-5.5 (High)
⚡ GPT-5.5 (Extra High)

Interactive plan selector

Plan display supports both automatic detection and manual override.

• Auto Detect uses session telemetry, in-memory cache, and persisted cache fallback.
• Manual mode writes the selected plan to ~/.codex/discord-presence-config.json.
• The selector is designed for live correction when Codex telemetry says Unknown, lags behind, or you simply want Discord/TUI to reflect the plan you actually use.

Example Discord state

A typical compact presence state can look like:

⚡ GPT-5.5 (Extra High) | Pro ($200/month) • $7.13 • 31.5M tok • Ctx 64% left • 5h 100% • 7d 100%

Surface Routing

Detection priority:

1. session_meta.originator contains desktop
2. fallback: session_meta.source contains desktop
3. otherwise: Codex CLI / Codex VS Code Extension profile

Command Reference

Configuration

Config file location:

• ~/.codex/discord-presence-config.json

Essential defaults:

Example:

{
  "schema_version": 8,
  "discord_client_id": "1470480085453770854",
  "discord_client_id_desktop": "1478395304624652345",
  "poll_interval_seconds": 2,
  "privacy": {
    "show_cost": true
  },
  "openai_plan": {
    "mode": "auto",
    "tier": "pro",
    "show_price": true
  }
}

Plan and model display notes:

• openai_plan.mode = "manual" makes openai_plan.tier the displayed account plan.
• openai_plan.mode = "auto" keeps telemetry and cache-based plan detection enabled.
• The TUI plan selector writes the same config file used at startup.
• Fast mode is derived from ~/.codex/.codex-global-state.json.
• Reasoning effort is derived from live turn_context telemetry.

Environment overrides:

• CODEX_DISCORD_CLIENT_ID
• CODEX_DISCORD_CLIENT_ID_DESKTOP
• CODEX_PRESENCE_STALE_SECONDS
• CODEX_PRESENCE_POLL_SECONDS
• CODEX_PRESENCE_ACTIVE_STICKY_SECONDS
• CODEX_HOME

Build and Artifacts

Published binaries are available in GitHub Releases.

Local Cargo build cache is stored under .build/target and final release binaries are copied into releases/<platform>/.

Expected artifact layout:

• releases/windows/codex-discord-rich-presence.exe
• releases/linux/codex-discord-rich-presence
• releases/macos/codex-discord-rich-presence
• releases/macos/codex-discord-rich-presence-x64
• releases/macos/codex-discord-rich-presence-arm64

Windows executable icon source:

• assets/branding/codex-app.png

Documentation

• Docs Index
• CLI and Presence Contract
• Local Data and Schema Contracts
• TUI Information Architecture

Security and Privacy

• Reads local Codex session files only.
• Uses Discord local IPC for activity publishing.
• Does not add an external telemetry backend.
• See PRIVACY.md and SECURITY.md.

Credits

OpenAI Brand Note

OpenAI marks and logos are trademarks of OpenAI.
Follow official brand policy: https://openai.com/brand/

License

MIT (LICENSE)