Admin & AI tools

The control room. Use the AI research assistant to discover and draft new places, then review the queue. These tools run against a local agent server on your machine — see “Run it yourself” below.

Published

In review

Provinces

AI-tagged

🤖 AI research assistant

Not connected

Agent server Provider Model

Couldn't reach the agent server. Start it locally with npm run agent:server (see “Run it yourself”). The public site can't reach localhost; run the site locally with npm run dev for the full workflow.

Review queue (0)

Items awaiting moderation. To approve, change approval: pending to published in the file (or use the agent's save step).

Queue is empty. 🎉

Moderation

Removing content = delete the file or revert the commit. Everything is version-controlled.

Run it yourself

cp .env.example .env   # add provider keys (optional)
npm run agent:server   # start the local agent API (:8787)
npm run dev            # start the site, then Connect above

# or one-shot from the CLI:
npm run agent -- research "giant statues in the south"

Works with Ollama (local, free), OpenAI, Anthropic, or Gemini. See docs/AI_AGENTS.md.