> ## Content Index
> Fetch the complete content index at: https://www.npiedy.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Project Summer: Query Your Letterboxd Data with SQL and Claude
- URL: https://www.npiedy.com/project-summer-query-your-letterboxd-data-with-sql-and-claude/
- Published: 2026-10-09T04:30:00.000Z
- Updated: 2026-10-09T10:56:49.000Z
- Description: Letterboxd is great, but it couldn’t answer my own questions. So I built Project Summer: a CLI and MCP server that turns my Letterboxd export into a DuckDB database I can query with SQL or just ask Claude about.
- Author: Naveen Piedy
- Tags: letterboxd, AI, Movies, Claude Dashboards, mcp, DuckDB

Letterboxd quickly rose to become the gold standard. So much so that plenty of people have googled "Letterboxd for X". Hell, I have been looking for a Letterboxd for video games forever. The platform was able to build all of this because movies have a well-maintained public database in the form of TMDB, which a lot of other domains lack.  
  
It is great for tracking what you watch, and Pro gives you insights and stats. But as someone who works with code and data on a daily basis, I wanted something more flexible. For instance, I wanted to compare the average number of days between a movie's release and when I watched it, for Kollywood versus Hollywood.  
  
So I did what a data engineer would naturally do. I built Project Summer, a tool for querying and exploring my own data. And it has a few more bells and whistles.

## How Project Summer Fits Together

One of the biggest features of Letterboxd is that it lets you export your data. You can get your lists, your diary, your reviews, everything in nice CSV files. Will the new owners want to keep that? Time will tell.

![](https://storage.ghost.io/c/f7/10/f7101a80-6932-4b03-b4dc-971108e4b353/content/images/2026/10/image.png)

Architecture Diagram

As the diagram above shows, you can ingest all that data into a DuckDB file, resolve it to the matching TMDB id, and then enrich it. Since it is in DuckDB, you can go nuts and query it however you see fit.  
  
But I wanted to go one step further. I designed a plugin system. Any repeatable task or tool can be registered by writing a single decorated function. I built a rank tool that puts movies head to head in a draft-style ranking for me to pick. And I built a query-to-list tool.  
  
Once registered, these tools can be used manually from the CLI or, if you want, by agents through MCP (a standard way for AI assistants to call tools), depending on your AI preference.

![](https://storage.ghost.io/c/f7/10/f7101a80-6932-4b03-b4dc-971108e4b353/content/images/2026/10/image-1.png)

CLI Interface of Project Summer

## DuckDB: Ask Your Data Anything

> **What is DuckDB?** A fast, in-process SQL database built for analytics. Like SQLite, it runs inside your program and keeps everything in a single file, with no server to set up. Unlike SQLite, it is designed for questions that scan lots of rows, such as counts, averages and group-bys.

With all the data inside a DuckDB file, we can query it and create SQL notebooks. DuckDB is something I discovered recently, and I have been head over heels in love with this tool. It is fast, works with a lot of data formats and, best of all, has a built-in UI.

![](https://storage.ghost.io/c/f7/10/f7101a80-6932-4b03-b4dc-971108e4b353/content/images/2026/10/image-2.png)

DuckDB UI

My assumption was correct that Kollywood movies are faster to get onto streaming. The median being 38 days compared to 56 points to that.

## Let AI Do the Asking

The other way to use it is to ask your agents, and they can query the data and come up with the answer. Here is an example of Claude pulling my data, querying it, and presenting the results.

![](https://storage.ghost.io/c/f7/10/f7101a80-6932-4b03-b4dc-971108e4b353/content/images/2026/10/image-4.png)

Claude in Action

  
Claude spotted it: I love Wes Anderson movies. Going further, you can even have Claude or another agent build you a live dashboard. Here is one it built. Everything runs read-only, so Claude can look at your data but never change it.

![](https://storage.ghost.io/c/f7/10/f7101a80-6932-4b03-b4dc-971108e4b353/content/images/2026/10/image-5.png)

Claude Dashboard

## The Fun Part: Tiny Silly Features

One of the features I built in is a random movie picker. I can give it a genre or max-runtime and it will return a movie from my watchlist. I also have a list builder, a movie ranker, and some other silly things. Each one is just a single decorated function, which is why adding them is so easy.

![](https://storage.ghost.io/c/f7/10/f7101a80-6932-4b03-b4dc-971108e4b353/content/images/2026/10/image-8.png)

Random Movie Picker from Watchlist

  
In fact, one of the upcoming plugins I want to build is a six degrees of separation game using actors from my own dataset: pick two actors, and it finds the shortest chain of films that connects them. The repo is open source, so anyone can fork it and add their own plugins.

## Why I Want to Own My Data

The chatter on the internet is that Letterboxd is for sale, with Netflix, Sony, A24 and the New York Times reportedly circling. I don't think they are going to sell it to me for 50 bucks. But the reason Letterboxd has stood out is that it hasn't been swallowed by a big studio or tech company. It has that passion project feeling.  
  
And time and again, we have seen companies swoop in and buy a beloved tool, site, etc., and enshittification ensues quickly. That is also a reason I want to own my data before it is too late. Thankfully, Letterboxd lets you export your data, and Summer is built on that export.

## Go Play With Your Own Data

Feel free to use [Project Summer](https://github.com/naveenpiedy/projectsummer?ref=npiedy.com) with your own Letterboxd export. If you hit a bug, please raise an issue on GitHub and I will take a look. If you have an idea for a new plugin, or a way to make Summer easier to extend, I would love to hear it. Your data is yours, so go have some fun with it!

If you enjoyed this, I'm hoping to write about projects like it fairly often, so please subscribe if you want them in your inbox. Thoughts, ideas for new plugins, or your own Letterboxd stats? Drop a comment, I read them all. And if you want to find me elsewhere: [goodreads](https://www.goodreads.com/user/show/120692883-naveen-piedy?ref=npiedy.com), [letterboxd](https://letterboxd.com/naveenpiedy/?ref=npiedy.com), [github](https://github.com/naveenpiedy?ref=npiedy.com) and [linkedin](https://www.linkedin.com/in/naveen-piedy-47b7336b/?ref=npiedy.com).