diff --git a/README.md b/README.md
index a49b0d4..f1b9b0c 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,8 @@
 # datafusion-forecast
 
+[![CI](https://github.com/nfnt/datafusion-forecast/actions/workflows/check.yml/badge.svg)](https://github.com/nfnt/datafusion-forecast/actions/workflows/check.yml)
+[![license](https://img.shields.io/github/license/nfnt/datafusion-forecast.svg)](LICENSE)
+
 Forecasts with Mistral AI's [Timestral](https://github.com/mistralai/timestral) time-series models inside
 [Apache DataFusion](https://datafusion.apache.org) SQL. The `forecast` aggregate function treats each group as a series
 and returns a median and 21 quantiles for each step ahead. The model runs natively in Rust on
diff --git a/crates/candle-timestral/README.md b/crates/candle-timestral/README.md
index 216a91c..cb5231d 100644
--- a/crates/candle-timestral/README.md
+++ b/crates/candle-timestral/README.md
@@ -1,5 +1,11 @@
 # candle-timestral
 
+[![crates.io](https://img.shields.io/crates/v/candle-timestral.svg)](https://crates.io/crates/candle-timestral)
+[![docs.rs](https://img.shields.io/docsrs/candle-timestral)](https://docs.rs/candle-timestral)
+[![CI](https://github.com/nfnt/datafusion-forecast/actions/workflows/check.yml/badge.svg)](https://github.com/nfnt/datafusion-forecast/actions/workflows/check.yml)
+[![license](https://img.shields.io/crates/l/candle-timestral.svg)](https://github.com/nfnt/datafusion-forecast/blob/main/LICENSE)
+[![MSRV](https://img.shields.io/crates/msrv/candle-timestral.svg)](https://crates.io/crates/candle-timestral)
+
 Inference for Timestral checkpoints (e.g. [`mistralai/Timestral-5M`](https://huggingface.co/mistralai/Timestral-5M))
 in Rust with [Candle](https://github.com/huggingface/candle), ported from the reference Python package. It loads the
 released `config.json` + `model.safetensors` and reproduces `TimestralForecaster.forecast`: multivariate context,
diff --git a/crates/datafusion-forecast/README.md b/crates/datafusion-forecast/README.md
index 7fb53ab..aefe4e7 100644
--- a/crates/datafusion-forecast/README.md
+++ b/crates/datafusion-forecast/README.md
@@ -1,5 +1,11 @@
 # datafusion-forecast
 
+[![crates.io](https://img.shields.io/crates/v/datafusion-forecast.svg)](https://crates.io/crates/datafusion-forecast)
+[![docs.rs](https://img.shields.io/docsrs/datafusion-forecast)](https://docs.rs/datafusion-forecast)
+[![CI](https://github.com/nfnt/datafusion-forecast/actions/workflows/check.yml/badge.svg)](https://github.com/nfnt/datafusion-forecast/actions/workflows/check.yml)
+[![license](https://img.shields.io/crates/l/datafusion-forecast.svg)](https://github.com/nfnt/datafusion-forecast/blob/main/LICENSE)
+[![MSRV](https://img.shields.io/crates/msrv/datafusion-forecast.svg)](https://crates.io/crates/datafusion-forecast)
+
 Forecasts with Mistral AI's [Timestral](https://github.com/mistralai/timestral) models as an
 [Apache DataFusion](https://datafusion.apache.org) aggregate function. The model runs in Rust on top of the
 [`candle-timestral`](https://github.com/nfnt/datafusion-forecast/tree/main/crates/candle-timestral) crate.