Connect Spritory to ChatGPT and Codex, OpenAI's coding agent, and ask for pixel art sprites and effects that match your game. One setup covers the ChatGPT desktop app, the Codex CLI and the Codex IDE extension, because they share the same settings file.
A Spritory token and the server address https://spritory.com/api/mcp. See how to connect to create a token. Below, replace spt_your_token with yours.
Set up with the Codex CLI
Codex reads your token from an environment variable, so it never sits in a config file. First add the server:
codex mcp add spritory --url https://spritory.com/api/mcp --bearer-token-env-var SPRITORY_TOKENThen set the variable in your shell profile, such as ~/.zshrc or ~/.bashrc, and open a new terminal:
export SPRITORY_TOKEN="spt_your_token"Set up in the config file
You can also add the server by hand. Open ~/.codex/config.toml and add:
[mcp_servers.spritory]
url = "https://spritory.com/api/mcp"
bearer_token_env_var = "SPRITORY_TOKEN"Set the environment variable as shown above.
Set up in the ChatGPT desktop app
- Open Settings, then MCP servers.
- Click Add server, name it
spritory, and choose Streamable HTTP. - Enter
https://spritory.com/api/mcpas the URL. - Save, then click Restart.
The desktop app reads the same config.toml. If the form has no field for the token, add the server with the Codex CLI or the config file instead, and it will show up in the app too.
The IDE extension
The Codex extension for VS Code and other editors uses the same settings. Once Spritory is in config.toml, the extension can use it.
ChatGPT in the browser only reaches MCP servers through installed plugins, so this setup does not apply there. Use the desktop app, the Codex CLI or the IDE extension.
Check the connection
codex mcp listSpritory should appear in the list. Start a new Codex session so it picks up the change.
Next steps
Ask your agent to make something in Spritory. For ideas, see prompt ideas. If something goes wrong, see troubleshooting.