This article contains affiliate links. We may earn a commission at no extra cost to you if you sign up through our links.
OpenAI Codex Just Got Plugins: Slack, Figma, Notion — But Happycapy Already Does This (and More)
OpenAI launched plugins for Codex on March 26, 2026 — adding Slack, Figma, Notion, Gmail, and Google Drive integrations for developers. It is a strong move for coding teams. But if you are not a developer, Happycapy already does all of this, plus memory, email delivery, and Mac control — with zero setup.
OpenAI's Codex just became a lot more useful. On March 26, 2026, the company shipped a first-class plugin system for its AI coding agent, letting developers connect it directly to Slack, Figma, Notion, Gmail, and Google Drive. No more manual credential wrangling or context-switching. Plugins bundle authentication, prompt workflows, and MCP server configs into a single installable package.
It is a significant upgrade. Codex already had 1.6 million weekly active users as of early March 2026. Plugins push it from "AI that writes code" to "AI that participates in your whole dev workflow." That is a meaningful line to cross.
But here is the thing: Happycapy has had a skills system for months. And it is built for everyone — not just developers. Let's break down what Codex plugins actually do, where they fall short, and why Happycapy's approach is different.
What OpenAI Codex Plugins Actually Do
Codex plugins ship with version 0.117.0. The system works across the Codex desktop app, the CLI, and IDE extensions for VS Code and JetBrains. Here is what the initial release includes:
- Slack — read channel history, post messages, search threads
- Figma — read design files and component properties directly in context
- Notion — pull docs and database records into Codex prompts
- Gmail — read emails, compose drafts, search inbox
- Google Drive — access files and folders for context
- 20+ more in the curated directory, with self-serve publishing coming soon
The plugin manifest lives in a .codex-plugin/plugin.json file. Developers can scaffold new plugins locally using the built-in @plugin-creator skill. Plugins sync per-repo or per-user via marketplace.json files — a clean, Git-native pattern.
The underlying model is GPT-5.3-Codex, which runs 25% faster than its predecessor. For engineering teams already in the OpenAI ecosystem, this is compelling.
The Developer Ceiling Problem
Codex plugins are, by design, a developer product. To use them you need to install the CLI, configure a marketplace.json, authenticate via device-code sign-in, and understand how MCP server configs work. That is reasonable for a senior engineer. It is a wall for everyone else.
Most people who want AI to "connect to Slack and summarize what I missed" or "pull my Gmail and draft replies" are not developers. They are founders, marketers, writers, and operators. For them, Codex plugins do not apply.
Codex Plugins vs. Happycapy Skills: Full Comparison
| Feature | OpenAI Codex Plugins | Happycapy Skills |
|---|---|---|
| Launch date | March 26, 2026 | Available since 2025 |
| Target user | Developers (CLI / IDE) | Anyone — no coding required |
| Slack integration | Yes (plugin) | Yes (skill) |
| Gmail / email | Read + draft | Full send via Capymail (your own alias) |
| Figma integration | Yes | Not natively (use browser skill) |
| Memory across sessions | No | Yes — persistent memory, MEMORY.md |
| Mac computer control | No | Yes — Mac Bridge (remote terminal + files) |
| Web search | Via tool calls | Built-in web search skill |
| Image generation | No | Yes (FLUX, Gemini, Grok, Seedream) |
| Social media scheduling | No | Yes (skill) |
| Setup complexity | CLI, JSON manifest, MCP config | One-click install from skills library |
| Works for coding | Yes (core use case) | Yes (Claude Code skills) |
| Pricing | ChatGPT Pro ($20/mo) + Codex access | Free / Pro $17/mo / Max $167/mo |
Want skills that actually work for non-developers? Happycapy has Slack, email, web search, memory, and Mac control — all in one agent. No CLI required.
Try Happycapy Free →Where Codex Plugins Shine
To be fair: for developer teams already using Codex, plugins are a genuine workflow upgrade. The Figma integration alone — reading design specs directly into a code generation prompt — removes real friction. The Notion integration means Codex can read your PRD and write code against it without copy-pasting. That is useful.
The plugin manifest pattern is also clean for teams. Checking a marketplace.json into a repo means everyone on the team gets the same Codex configuration automatically. That is good ops practice.
Where Happycapy Wins
Happycapy's skills system was built for people, not pipelines. Here is what you get that Codex plugins do not offer:
- Persistent memory — Happycapy remembers your preferences, projects, and context across every session. Codex starts fresh each time.
- Capymail — Happycapy can actually send emails from your personalized alias. Codex reads and drafts but does not send.
- Mac Bridge — Control your Mac remotely, run terminal commands, and read/write files via Happycapy's Mac Bridge. Codex has no Mac control layer.
- No coding required — Install a skill in one click. No JSON, no CLI, no manifest files.
- Full-stack agent — Happycapy handles writing, research, social media, image generation, and coding in one place. Codex is a coding tool with integrations.
Who Should Use Which
Use Codex plugins if: you are an engineer who already lives in Codex, loves CLI workflows, and wants to pull Figma or Notion context into your coding sessions.
Use Happycapy if: you want an AI agent that handles your whole workday — email, writing, research, social media, and code — with memory, no setup friction, and skills that actually work for non-technical users.
The two tools serve different people. Codex plugins are powerful in the right hands. Happycapy is powerful for everyone.
One agent for everything. Happycapy Pro is $17/month and includes email, memory, Mac Bridge, web search, and 50+ skills. No coding required.
Start Free on Happycapy →Frequently Asked Questions
OpenAI Codex plugins are installable packages that connect Codex to external tools like Slack, Figma, Notion, Gmail, and Google Drive. They launched on March 26, 2026, and work across the Codex desktop app, CLI, and IDE extensions for VS Code and JetBrains.
Both extend an AI agent with tool integrations. Codex plugins are developer-centric and require CLI or IDE setup. Happycapy skills work for any user with one-click installs, and Happycapy also adds persistent memory, email sending, and Mac control that Codex does not have.
No. Codex plugins are designed for developers using the CLI and IDE extensions. If you are not a developer, Happycapy's no-code skills system is a better fit for connecting AI to your tools.
Happycapy is a full AI agent platform that includes Claude Code skills for developers, but goes beyond coding. It handles writing, research, email, social media, Mac control, and memory in one subscription. Codex is focused exclusively on coding workflows.
Sources: OpenAI Developers changelog (developers.openai.com), Neowin (March 26, 2026), Releasebot.io (March 27, 2026), ZDNET (March 27, 2026).
Comments are coming soon.