Run AI locally, nothing leaves my server
Client work, private drafts, or you just do not want to pay per token. WordPress can talk to a model running on your own machine. You need a provider plugin that points the built-in AI Client at it.
-
The simplest route. Run Ollama, install this, and any plugin built on the WordPress AI Client uses your local model.
By Fueled. Run local open-source models with zero data leaving the server.
-
Use this if you run LM Studio or anything else that speaks the OpenAI API. One plugin, lots of endpoints.
Connects the WordPress 7.0 AI Client to any OpenAI-compatible endpoint (Ollama, LM Studio, OpenRouter, Groq, and ~110 presets).
-
Only if Open WebUI is already your setup. It finds your models automatically.
OpenAI-compatible settings with automatic model discovery.
-
A different kind of local: let a coding agent work on a local copy of your site instead of the live one.
Automattic's local dev app exposes an MCP server (Settings → MCP) so agents like Claude Code and Copilot can create, start, and manage local sites, edit files, query the database, and run WP-CLI.
These are picks, not rankings. Think something is missing or wrong? Tell me.