diff --git a/.github/workflows/pkgdown.yaml b/.github/workflows/pkgdown.yaml
index c9f0165..be08a0a 100644
--- a/.github/workflows/pkgdown.yaml
+++ b/.github/workflows/pkgdown.yaml
@@ -37,6 +37,20 @@ jobs:
           extra-packages: any::pkgdown, local::.
           needs: website
 
+      - name: Cache Spark
+        id: cache-spark
+        uses: actions/cache@v2
+        with:
+          path: /home/runner/spark/spark-3.3.4-bin-hadoop3
+          key: sparklyr-spark-3.3.4-bin-hadoop3-3
+
+      - name: Install Spark (via sparklyr)
+        if: steps.cache-spark.outputs.cache-hit != 'true'
+        run: |
+          sparklyr::spark_install(
+            version = Sys.getenv("3.3.4"),
+            hadoop_version = Sys.getenv("3")
+
       - name: Build site
         run: pkgdown::build_site_github_pages(new_process = FALSE, install = FALSE)
         shell: Rscript {0}
diff --git a/vignettes/articles/databases.Rmd b/vignettes/articles/databases.Rmd
new file mode 100644
index 0000000..e3f4e53
--- /dev/null
+++ b/vignettes/articles/databases.Rmd
@@ -0,0 +1,103 @@
+---
+title: "Using Databases"
+---
+
+```{r, include = FALSE}
+knitr::opts_chunk$set(
+  collapse = TRUE,
+  comment = "#>",
+  message = FALSE
+)
+```
+
+## Setup
+
+We start by loading our packages and creating a simple fitted workflow using `mtcars`.
+
+```{r setup}
+library(orbital)
+library(recipes)
+library(parsnip)
+library(workflows)
+```
+
+```{r}
+rec_spec <- recipe(mpg ~ disp, data = mtcars) |>
+  step_impute_mean(all_numeric_predictors()) |>
+  step_normalize(all_numeric_predictors())
+
+lm_spec <- linear_reg()
+
+wf_spec <- workflow(rec_spec, lm_spec)
+wf_fit <- fit(wf_spec, data = mtcars)
+```
+
+then create our orbital object.
+
+```{r}
+orbital_obj <- orbital(wf_fit)
+orbital_obj
+```
+
+## SQL
+
+```{r}
+library(DBI)
+library(RSQLite)
+
+con_sqlite <- dbConnect(SQLite(), path = ":memory:")
+mtcars_sqlite <- copy_to(con_sqlite, mtcars, name = "mtcars_table")
+
+predict(orbital_obj, mtcars_sqlite)
+```
+
+```{r}
+predict(orbital_obj, mtcars_sqlite) |>
+  collect()
+```
+
+## Spark
+
+```{r}
+library(sparklyr)
+
+con_spark <- spark_connect(master = "local")
+mtcars_spark <- copy_to(con_spark, mtcars, overwrite = TRUE)
+
+predict(orbital_obj, mtcars_spark)
+```
+
+```{r}
+predict(orbital_obj, mtcars_spark) |>
+  collect()
+```
+
+## duckdb
+
+```{r}
+library(duckdb)
+
+con_duckdb <- dbConnect(duckdb(dbdir = ":memory:"))
+mtcars_duckdb <- dplyr::copy_to(con_duckdb, mtcars)
+
+predict(orbital_obj, mtcars_duckdb)
+```
+
+```{r}
+predict(orbital_obj, mtcars_duckdb) |>
+  collect()
+```
+
+## arrow
+
+```{r}
+library(arrow)
+mtcars_arrow <- as_arrow_table(mtcars)
+
+predict(orbital_obj, mtcars_arrow)
+```
+
+```{r}
+predict(orbital_obj, mtcars_arrow) |>
+  collect()
+```
diff --git a/vignettes/supported-models.Rmd b/vignettes/supported-models.Rmd
index dfa6166..ba37e69 100644
--- a/vignettes/supported-models.Rmd
+++ b/vignettes/supported-models.Rmd
@@ -2,7 +2,7 @@
 title: "Supported Models and recipes steps"
 output: rmarkdown::html_vignette
 vignette: >
-  %\VignetteIndexEntry{supported-models}
+  %\VignetteIndexEntry{Supported Models and recipes steps}
   %\VignetteEngine{knitr::rmarkdown}
   %\VignetteEncoding{UTF-8}
 ---