Codex Discord Rich Presence

First-class pricing, model label, and ⚡ Fast mode visibility for both GPT-5.5 and GPT-5.5 Pro — official OpenAI rates applied the moment Codex switches models, no config touch required.

Local-first by design: the runtime reads your Codex session files on disk, talks to Discord over local IPC. Nothing is stored in the cloud.

Overview

Codex Discord Rich Presence reads local Codex session telemetry, detects the active Codex surface, renders a live terminal dashboard, and publishes a clean Discord presence with activity, model, Fast mode, reasoning effort, account plan, token usage, cost, context window, and quota visibility.

What it shows

Signal	Example	Notes
Surface	Codex App / Codex CLI	Auto-routed from session_meta.originator and session_meta.source
Activity	Reading session.rs	Derived from response items, tool calls, and commentary signals
Model	⚡ GPT-5.5 (Extra High)	Fast mode adds the lightning prefix; reasoning effort adds the suffix
Plan	Pro ($200/month)	Auto-detected from telemetry or forced manually from the TUI/config
Live context	Ctx 64% left	Derived from active-turn usage and model context window
Limits	5h 100% • 7d 100%	Tracks global account quota freshness and remaining percentages
Cost	$7.13	Uses pricing aliases and overrides from local config

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

Active Surface	Discord App Profile	Client ID Field	Main Large Asset
Codex CLI / Codex VS Code Extension	Codex	discord_client_id	codex-logo
Codex App Desktop	Codex App	discord_client_id_desktop	codex-app

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

Command	Purpose
codex-discord-presence	Starts the runtime, TUI, and Discord IPC loop
codex-discord-presence codex [args...]	Runs as a Codex passthrough wrapper while keeping presence active
codex-discord-presence status	Prints health, active session state, limits source, model label, plan, Fast mode, and reasoning effort
codex-discord-presence doctor	Runs diagnostics for session roots, config, Discord IDs, and command availability

Configuration

Config file location:

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

Essential defaults:

Key	Value
schema_version	8
discord_client_id	1470480085453770854
discord_client_id_desktop	1478395304624652345
display.large_image_key	codex-logo
display.desktop_large_image_key	codex-app
display.desktop_large_text	Codex App
display.small_image_key	openai
privacy.show_cost	true
openai_plan.mode	auto
openai_plan.tier	pro
openai_plan.show_price	true
poll_interval_seconds	2

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)