ChatGPT plugins for Development
695 plugins, apps and MCP servers in Development, ranked by official status, traction and completeness. Page 1 of 12.
Developers mostly use ChatGPT plugins to bring the rest of their toolchain into a coding session: reading pull requests and CI logs, pulling current library docs, driving a browser to check a fix, and scanning a diff before it ships. Most of this happens in Codex, which shares one Plugin Directory with ChatGPT, so a plugin you find in one is also listed in the other.
Three kinds of listing appear here. A plugin is the package you install from the directory; developer plugins usually bundle an app connection with skills that tell the model how to work with it, such as how to triage CI or which Stripe API to prefer. An app is the connection on its own. An MCP server is the underlying service, which you add yourself: in Codex with codex mcp add, or in ChatGPT through developer mode. Local (stdio) servers such as Chrome DevTools run on your machine, so they work in Codex, not in ChatGPT on the web.
Before you connect anything, check what it can write. Many developer tools can open pull requests, merge, deploy or run scripts. Prefer read-only modes or narrow scopes where a service offers them, keep confirmations on for write tools, and connect test projects first. Some plugins are marked Desktop only and need the ChatGPT desktop app, developer mode is limited to Plus, Pro, Business, Enterprise and Education accounts, and in a company workspace your admin decides which plugins you can install. For a longer list with setup notes, see the best ChatGPT plugins for developers.
Top picks for Development
OpenAI's GitHub plugin pairs the GitHub connection with a CLI workflow for triaging pull requests and issues, debugging failing CI and publishing changes. It is the default pick for anyone whose code lives on GitHub. If you want to choose exactly which tools are exposed, GitHub's own MCP server lets you enable toolsets selectively.
Context7, from Upstash, gives coding agents current, version-specific documentation and examples for thousands of libraries, so they stop reaching for APIs that changed two releases ago. It runs as a hosted server or a local npm package and needs a Context7 API key.
Google's Chrome DevTools server lets an agent control a live Chrome instance: read console errors and network requests, record performance traces and run Lighthouse audits. It suits front-end debugging and runs locally, so use it from Codex. Microsoft's Playwright MCP is the alternative when you need Chromium, Firefox and WebKit.
OpenAI's Codex Security plugin bundles skills for repository and diff scans, threat models, finding triage, fix proposals and verification, and vulnerability write-ups. Run a diff scan on a branch before you open a pull request, and have it track findings in Linear, Jira or GitHub issues.
An OpenAI skills plugin for front-end work: turning image concepts into apps, browser-based testing and debugging, React and Next.js performance, shadcn/ui, Stripe integration and Supabase Postgres. It suits anyone starting a web app in Codex. It contains skills only, so pair it with the Stripe or Supabase plugin when you need a live account connection.
All 695 in Development
SuperpowersPlan, develop, and debug codeplugin · by Jesse Vincent · 294k stars
SerenaSemantic code retrieval and editing toolkit for coding agentsMCP server · by Oraios AI · 30k stars
Task MasterAI-powered task management from PRDs for coding agentsMCP server · by Eyal Toledano · 28k stars
ShopifyBuild and manage your storeplugin · by Shopify · 7.2k starsDevelopment plugins: common questions
- Which ChatGPT plugins work with GitHub?
There are three options. The GitHub plugin, built by OpenAI, combines the GitHub connection with workflows for pull requests, issues and CI. The GitHub app is OpenAI's first-party connection on its own; OpenAI says it is required for some features, including Codex. GitHub's MCP server is for clients where you configure servers yourself: it offers a hosted OAuth endpoint or a local server with a personal access token.
- Do developer plugins work in ChatGPT or only in Codex?
Mostly both. ChatGPT and Codex share one Plugin Directory, so a public plugin is listed in each, though some capabilities only work on one surface. Plugins marked Desktop only need the ChatGPT desktop app, and the Codex IDE extension does not support plugins. Local MCP servers that run as a command on your machine, such as Chrome DevTools or Playwright, need Codex. Setup: Use plugins in Codex.
- How do I add an MCP server that has no plugin?
In Codex, run
codex mcp add <name> -- <command>for a local server, or pass--urlfor a hosted one;codex mcp listshows what is configured. In ChatGPT, turn on developer mode under Settings → Security and login, then add the server's URL from the Plugins page. Developer mode is available on Plus, Pro, Business, Enterprise and Education accounts on the web. Full walkthrough: Add an MCP server to ChatGPT.- Is it safe to give ChatGPT access to my repositories?
It can be, with limits. Give a plugin only the scopes it needs, keep confirmations on for write actions such as merging or deploying, and connect test projects first. Take extra care when an agent reads untrusted content, such as issues on a public repo, in the same session where it can push code: that is where prompt injection does damage. Our security checklist covers each step.
Guide updated Oct 24, 2026.