LM Studio
Time: about one command · 1 min.
Copy one line
Section titled “Copy one line”Paste this into a terminal where the tool is already set up. It adds the connection; nothing else changes.
# LM Studio → Program → mcp.json{"mcpServers": {"palaia": {"type": "streamable-http", "url": "http://palaia.local:8420/mcp/default", "headers": {"Authorization": "Bearer <paste-your-token>"}}}}Replace <paste-your-token> with the token the dashboard shows when you click Issue token on its connect page — it is shown once, so copy it then. Every request needs it; without it the hub turns the tool away.
Or just ask it
Section titled “Or just ask it”If you would rather not touch a terminal, paste this to the AI itself and let it set itself up:
Please connect yourself to my palaia hub as an MCP server:http://palaia.local:8420/mcp/defaultSend the header "Authorization: Bearer <paste-your-token>" with every request.Then run a test recall and tell me what you found.Same here: replace <paste-your-token> with your token.
Or save a file
Section titled “Or save a file”Some setups read this from a file instead of a command. Save it as palaia-lmstudio-mcp.json:
{ "mcpServers": { "palaia": { "type": "streamable-http", "url": "http://palaia.local:8420/mcp/default", "headers": { "Authorization": "Bearer <paste-your-token>" } } }}Same here: replace <paste-your-token> with your token.
Teach it to look things up and save things on its own
Section titled “Teach it to look things up and save things on its own”LM Studio is an MCP host, not a skill loader — it has no place to put a SKILL.md. The memory still works; put the same guidance in the model’s system prompt, or ask for it directly.
Check it worked
Section titled “Check it worked”Ask it to remember something, then ask a different connected AI whether it knows the same thing. If both answer the same way, the connection is live — see Your first shared memory for the full walkthrough.