Connect with us

Artificial Intelligence

OpenAI Adds Plugin Marketplace to Codex

mm

OpenAI has launched a plugin system for Codex, its AI coding agent, adding a curated directory of integrations that connect the tool to workplace apps including Slack, Notion, Figma, Gmail, and Google Drive. The update, shipped with Codex version 0.117.0, positions the product as a workflow platform that extends beyond pure code generation.

What the Plugin System Does

Plugins in Codex are installable bundles that can contain three components: skills (predefined prompt workflows that guide agent behavior), app integrations (connectors to external services), and MCP server configurations (remote tools or shared context). According to the official plugin documentation, the system is designed to “make it easier to share the same setup across projects or teams.”

Image: OpenAI

The plugins work across all Codex surfaces — the desktop app, command line interface, and IDE extensions. Developers can browse and install plugins from a curated directory inside the Codex app. Support for a local per-repo marketplace and a personal per-user marketplace is also available, using marketplace.json files to define installable packages. Self-serve publishing to the official directory is listed as coming soon.

A built-in @plugin-creator skill scaffolds new plugins locally, generating the required .codex-plugin/plugin.json manifest along with optional directories for skills, app connectors, and MCP configurations. This lets teams build internal plugins and test them before distributing across projects.

The Codex changelog for version 0.117.0 describes plugins as a “first-class workflow,” with the update also adding remote sync for plugin install state, a plugin suggestion allowlist, and an improved plugin menu with installation status sorting.

Expanding Codex Beyond Coding

OpenAI is positioning the plugin expansion as part of a broader push to make Codex useful outside of software development. The company says plugins also support planning, research, and team coordination tasks — extending the tool’s utility to non-engineering workflows.

This direction aligns with comments from Thibault Sottiaux, head of OpenAI’s Codex product, who described the tool earlier this month as “becoming the standard agent” for enterprise deployments. Speaking to Fortune, Sottiaux said Codex’s core training emphasizes “instruction following, understanding large amounts of data, finding its own context, and navigating the world in order to make decisions” — capabilities that translate beyond code.

Codex reached 1.6 million weekly active users as of early March 2026, more than tripling after the February launch of GPT-5.3 Codex. Weekly token usage grew fivefold over the same period. Enterprise customers including Cisco, Nvidia, Ramp, Rakuten, and Harvey have deployed the tool across developer teams.

The plugin push also continues a pattern of OpenAI deepening its relationship with Slack. OpenAI hired former Slack CEO Denise Dresser as its first Chief Revenue Officer earlier this year, and the Slack plugin for Codex now makes the integration functional at the agent layer. Anthropic’s Claude Code has also moved in this direction, adding native Slack support for in-chat development workflows.

Strategic Context

The plugin marketplace fits into a larger product roadmap for OpenAI’s developer tools. ChatGPT’s Atlas browser for macOS, launched with agent mode and memory, and ChatGPT’s expanded creative toolset show OpenAI building out a suite of interconnected products around a common agent infrastructure. Codex plugins use the same Skills architecture described in those products, suggesting the framework is becoming shared across the platform.

The open question is distribution and ecosystem density. A plugin marketplace only becomes a durable moat if developers actively build and publish tools for it — the same dynamic that determines the success of app stores, browser extensions, and IDE plugin ecosystems. OpenAI currently controls the curated directory and has not announced a revenue-sharing model for third-party plugin developers. Whether the self-publishing feature, once released, attracts a meaningful developer ecosystem will determine how central the marketplace becomes to Codex’s value proposition.

Alex McFarland is an AI journalist and writer exploring the latest developments in artificial intelligence. He has collaborated with numerous AI startups and publications worldwide.