OpenAI Codex Integration
Codex + PostQuickAI = autonomous social media from your terminal. MCP server + 6 skills + AGENTS.md — all three open standards, all working together.
Requires Codex CLI and the PostQuickAI API add-on.
codex mcp add postquickai -- npx -y @postquickai/mcpThe -- separator is required — everything after it is the command and its args.
Then edit ~/.codex/config.toml to add your API key env var:
[mcp_servers.postquickai]
command = "npx"
args = ["-y", "@postquickai/mcp"]
env = { POSTQUICKAI_API_KEY = "pq_live_your_key_here" }Prefer a project-scoped config? Put it in .codex/config.toml in your repo root (trusted projects only).
Add opinionated workflows for common social media tasks:
# Plugin install (recommended)
codex plugin install postquickai/agent-skills
# Or manual
git clone https://github.com/kodenark/postquickai-agent-skills
cp -r postquickai-agent-skills/skills ~/.codex/skills/Catalog of skills: postquick.ai/agent-skills.
Codex treats AGENTS.md as its primary project-level config. Paste our snippet from /agents-md and any Codex session in that project understands when and how to use the PostQuickAI MCP tools.
In a Codex session:
MCP server, Skills, and REST API are all included with the API add-on on any paid plan.
Easiest: `codex mcp add postquickai -- npx -y @postquickai/mcp` then edit ~/.codex/config.toml to add the POSTQUICKAI_API_KEY env var. Or add the whole [mcp_servers.postquickai] block to config.toml directly.
Global: ~/.codex/config.toml. Project-scoped (trusted projects only): .codex/config.toml in your repo root. MCP servers are defined in [mcp_servers.<name>] tables.
Yes. Codex loads SKILL.md files from ~/.codex/skills/ (user-level) and .agents/skills/ (project-level). Run `codex plugin install postquickai/agent-skills` or clone github.com/kodenark/postquickai-agent-skills and copy the skills/ folder.
Yes — Codex treats AGENTS.md as its primary project-config file. Paste our snippet from /agents-md and Codex automatically understands how to use PostQuickAI in that project.
MCP tools are the low-level primitives (createPost, publishPost, getAnalytics). Skills are opinionated workflows that chain those tools (weekly-content-batch, analyze-post-performance). Install MCP first — Skills depend on it.
Yes. Use `codex exec "..."` with the MCP config loaded. Set POSTQUICKAI_API_KEY in the CI environment. Useful for nightly analytics digests, failed-post retry automations, etc.
Almost always a config format issue. Make sure command is just npx and args is an array — never put the whole command in one string. Also fully quit and restart your agent after editing config.
MCP and REST API access both require the API add-on on your plan. Enable it at postquick.ai/pricing.
Generate a key at postquick.ai/dashboard/api. Keys start with pq_live_. Put it in the env block of your MCP config.
Basic plan: 30 req/min, 5k/day. Pro plan: 120 req/min, 50k/day. See rate limits in the API docs.
Still stuck? See the developer docs or email support@postquick.ai.
PostQuickAI is agent-native — same API, same MCP server, same skills bundle across every major AI coding agent.
One CLI command and you're set. Use Codex for code, use it for social too.
Start free trial