diff --git a/README.md b/README.md
index 6005ed9..577f912 100644
--- a/README.md
+++ b/README.md
@@ -413,6 +413,7 @@ The DuckLake file format was released on 2025-05-27: see the [website](https://d
 - [SixSentences](https://github.com/SixSentences/sixsentences) - Self-hostable research workspace that compiles boolean literature-search queries to SQL and runs them offline with DuckDB over versioned Parquet snapshots of a scholarly corpus.
 - [GTFS Scorecard](https://github.com/ChelseaKR/gtfs-scorecard) - Data-quality grades for 2,400+ public transit (GTFS) feed records. DuckDB builds the cross-agency query layer and a public Parquet table that DuckDB can query over HTTP, and the site's SQL page runs DuckDB-WASM in the browser.
 - [PH Catalog](https://github.com/sarthakagrawal927/ph-catalog) - Resumable local Product Hunt catalogue and analytics pipeline that normalizes records into DuckDB and exports verified Parquet snapshots.
+- [AskCSV](https://github.com/Sahilll15/askcsv) - Ask questions about a CSV in plain English. The file is loaded into DuckDB-WASM in the browser, a model writes a read-only DuckDB query from the schema alone, and failed queries are repaired automatically.
 
 ## Integrations