@levelsio (Pieter Levels)

Use this blog inside Claude, Grok and ChatGPT

This blog is an MCP server: plug it into your AI assistant and it can read and search every post while you chat.

878Posts

98Tags

FreePrice

NoneSignup

MCP (Model Context Protocol) is the standard that lets AI assistants use external tools. Add the URL below to Claude, Grok, ChatGPT, Cursor or any MCP client, and when you ask about startups, AI, travel or nomad life it can pull the actual posts — titles, dates, tags and full text — instead of guessing.

https://levels.io/mcp

How to connect

Claude (web / desktop)

Settings → Connectors → Add custom connector
Name it levelsio, paste https://levels.io/mcp, no auth needed
Ask: “what does @levelsio say about being a digital nomad?”

Claude Code

claude mcp add --transport http levelsio https://levels.io/mcp

ChatGPT

Settings → Apps & Connectors (enable Developer mode if you don't see custom connectors)
Add connector with URL https://levels.io/mcp, no authentication

Grok

Open Grok on grok.com or in the 𝕏 app, then Settings → Connectors
Add an MCP server / custom connector
Name it levelsio, URL https://levels.io/mcp, transport HTTP (streamable), authentication none
Enable the connector, then ask: “using the levelsio connector, what does he say about bootstrapping vs raising VC?”

Naming the connector in your question helps Grok pick it up the first time. If you don't see custom connectors, the option is rolling out by account tier — the JSON API works from Grok's browsing tool in the meantime.

Other MCP clients

Add an MCP server with transport http (streamable HTTP) and URL https://levels.io/mcp. The server is stateless JSON-RPC 2.0 (protocol 2025-06-18); machine docs at /llms.txt and /api.

What your assistant can do

`search_posts`	Search @levelsio's blog posts by words in the title, tags or body text, optionally filtered by tag or date. Returns matching posts with title, date, tags and URL. Use get_post for the full text.
`get_post`	Get one blog post in full by its slug, as markdown plus metadata. Slugs come from search_posts.
`list_tags`	List every tag on the blog with how many posts use it.

Things to ask

“What does @levelsio actually argue about bootstrapping versus raising VC money?”
“Summarise everything he has written about living and working in Japan”
“He sold everything and flew to Bangkok in 2013 — what did he write at the time?”
“What is his case about air conditioning in Europe, and when did he start making it?”
“How did I grow Nomads.com, in my own words?”
“What has he posted about Photo AI revenue this year?”

Fair use

Free, no signup, no key. Rate limited to 5 requests a second per IP (short bursts of 20 are fine), shared with the JSON API — enough for any real conversation, not enough to bulk-scrape the archive. Prefer one /api/posts call over fetching every slug in turn. The writing is @levelsio's; please link back to the post if you quote it.

Read-only, published posts only — there are no write tools. Prefer plain HTTP? Use the JSON API, or llms.txt for the whole map. Made by @levelsio.