Codex

PublicPin — Lets people report local civic issues with a photo, category, and GPS location on a public map.

Spec Architect — Guides spec-driven development from proposal to requirements to tasks, has AI coding agents execute the tasks, and verifies results against the spec.

EnigmaGrid — Runs a volunteer distributed computation that searches for solutions to an unresolved historical Enigma message on users' Windows machines.

StoreFast — Publishes Microsoft Store app updates in one step, auto-translating release notes and forwarding certification and review alerts.

ExitEcho — Detects processes that keep running after an application's window is closed and lets the user terminate them.

Deskly — Lets users manage AI coding agents as employees in a 3D office, assigning them tasks in a project folder with approval before commands run.

ParrotVoice — Lets a computer act as a phone's Bluetooth headset so a voice AI can answer and place calls on the user's existing phone number.

Agora — Runs a shared chat channel where a human and multiple AI models collaborate on tasks, with web search, code execution, and image generation.

Jankcraft — Runs a World of Warcraft private server playable in a browser, with an AI agent harness that controls the game client over a websocket.

Omnesis — Indexes a person's personal data sources into a self-hosted graph and lets AI agents query it under per-agent access and privacy rules.

Chronicle — Runs a fantasy text RPG where an AI game master narrates while the engine enforces world rules, inventory, injuries and travel as database state.

Poker Double or Nothing — Lets players play a poker-based double-or-nothing game where they double credits by guessing the next card and winning the gamble phase.

Trace UI — Extracts UI components, design tokens and assets from any website and exports them as clean React, Tailwind or HTML code.

PR Bunny — Runs a local AI code review of GitHub pull requests and lets you pick which comments get posted as one review.

Astra — Generates short promotional trailer videos for a game by writing code that renders the animation.

GhosttyEXTREME — Shows the live status of multiple coding agent sessions in a terminal, with a file editor that follows agent edits, a repo map, and infrastructure views.

Vibe Multiagent — Orchestrates multiple AI coding agents with role separation, safety hooks that block risky commands, and an offline HTML dashboard for monitoring their tasks.

Token Harness — Installs and connects token-reducing optimizers to coding agents and routes suitable subtasks to cheaper models, shown in a local dashboard.

Talyn — Watches pull requests, sends an agent back to fix failed CI or merge conflicts, and merges PRs once they are green.

Orbtile — Shows local coding sessions on a Logitech keypad and a Mac menu-bar popover so users can see which sessions need attention and read their replies.

Gigamail — Manages email with an AI agent that drafts replies from a documents folder, schedules appointments, and sends follow-ups only after human approval.

TokenFish — Shows remaining Codex and Claude Code quota and reset times in a small animated desktop widget.

token-tv — Shows Codex and Claude usage limits on a cheap Wi-Fi desk clock display.

Updater — Records what coding agents ship, why, tradeoffs, files and expected impact, and makes that history searchable.

explainroo — Generates explainer videos from a topic by having a coding agent write the script and draw each scene in code, with a local voice.

cliproxy-rs — Exposes existing Claude, ChatGPT and Gemini subscription accounts as a single local API endpoint for coding tools.

OpenRod — Runs AI coding agents in isolated sandboxes with policy controls for network egress, secrets, and audit logging.

Codex on Crack — Orchestrates coding work across multiple AI models inside Codex, with a visual panel showing runs, reviews, and usage.

ADing — Puts multiple coding agents into a shared conversation where a leader agent can coordinate and delegate work.

drip — Estimates water usage of AI coding agent sessions and displays it live in the terminal status line or a meter beside the agent.

KV22Code — Lets standard ChatGPT Chat read, edit, and run commands on files, terminal, Git, and local dev tools on the user's computer through a local MCP connector.

OmniRush — Provides an AI agent harness for coding and other work, with inference costs covered for approved users.

Jive — Runs a terminal coding agent that executes tasks as DAG graphs of tool calls and fast small-model decisions instead of repeated LLM tool-call loops.

Routera — Routes AI model requests across providers from a single prepaid balance so users avoid per-subscription usage limits.

Zinley — Runs an AI agent that takes delegated goals across devices and tools, with its own phone number and email for outside contact.

Creativly Agent — Runs an AI agent that works on images, documents, notes and files laid out on a shared canvas.

I built a Tamagotchi for… — Displays a virtual pet-style companion on a physical device that reacts to a developer's coding activity.

Nodarama Verbatim — Records code copied from AI chat pages in a Chrome browser and writes it into your project files, so you don't have to paste it manually.

Scanify — Scans a codebase for security issues like exposed keys, open endpoints, and broken auth, and explains each in plain language with a fix.

If Codex Was 100% Free With… — Runs as a desktop app where an AI agent reads a project's files, writes and edits code across files, and executes terminal commands.

Agent OS — Orchestrates planning, sandboxed coding, build verification, and approval-gated deployment for long-term software projects using AI agents with persistent memory.

Nodus — Coordinates multiple specialized AI agents for planning, coding, review and testing, routing each task to a suited model.

MCPRelay — Runs a local MCP gateway that lets an MCP client such as ChatGPT execute actions on the user's computer.

Lamprey — Provides a desktop coding IDE that runs coding agents on DeepSeek, Gemma, Qwen, and OpenRouter models, with a planner-coder-reviewer agent pipeline, git diff review, and a terminal.

Row-Bot — Runs a local desktop AI workbench that routes requests across local and hosted models, with coding, memory, voice, workflows and tool integrations.