> For the complete documentation index, see [llms.txt](https://docs.mapbox.com/mapbox-tiling-service/llms.txt)

# Raster Recipe Details

A **recipe** is a JSON document composed of configuration options that tell Mapbox Tiling Service (MTS) how to turn a raster source data into MRT or raster tiles. Recipes allow you to specify details like:

-   The minimum and maximum zoom levels at which you want the features in your tilesets to be visible.
-   Which bands to use for processing.
-   Configure tilesize, buffer and scale for "rasterarray" sources.

Recipes must be constructed according to the rules in the [Recipe specification](https://docs.mapbox.com/mapbox-tiling-service/recipe-specification/).

To see example recipes for frequently-requested use cases, see the [Recipe examples page](https://docs.mapbox.com/mapbox-tiling-service/examples/).

You can use the [Data Workbench](https://docs.mapbox.com/console-tools/data-workbench/) to see the recipe for an existing tileset created with MTS. In the Mapbox Studio [Data Workbench](https://studio.mapbox.com/studio/data-workbench/), click on the tileset's name to open Data Workbench. Click on **Job history**, then click on **View recipe**.

## Common recipe filters

### Filtering bands based on their index

Recipe filters are quite powerful, but occasionally you might need a filter to target an individual band may not have any band metadata associated with it. In cases like this, we can use the `bandindex` filter. In this example, we're specifying a filter that uses the `bandindex` in the expression to filter the band at index 1.

```json
{
  "version": 1,
  "type": "rasterarray",
  "minzoom": 1,
  "maxzoom": 3,
  "sources": [
    {
      "uri": "mapbox://tileset-source/{ACCOUNT}/source_1"
    }
  ],
  "layers": {
    "basic": {
      "tilesize": 256,
      "resampling": "nearest",
      "buffer": 1,
      "source_rules": {
        "filter": ["all", ["in", ["bandindex"], ["literal", [1]]]]
      }
    }
  }
}
```

If we wanted to target multiple bands based on their indices, we would change the filter to include a list of all the indices we want. With this filter, we'd get bands 1 through 5.

```json
{
  "filter": ["all", ["in", ["bandindex"], ["literal", [1, 2, 3, 4, 5]]]]
}
```

#### Filtering bands based on a source tag and band index

Let's look at a slightly more complex recipe where we have two layers and want each layer to include data from a specific source file that does not include band metadata to aid in the filtering. In this scenario, we can use the `sourcetag` filter along with adding a `tag` property on a source. We're also including the `bandindex` in the filter here to make sure our layers have the appropriate bands from each file. This filter can be modified to filter bands on specific band metadata if needed.

```json
{
  "version": 1,
  "type": "rasterarray",
  "minzoom": 2,
  "maxzoom": 11,
  "sources": [
    {
      "uri": "mapbox://tileset-source/{ACCOUNT}/source_1",
      "tag": "source1"
    },
    {
      "uri": "mapbox://tileset-source/{ACCOUNT}/source_2",
      "tag": "source2"
    }
  ],
  "layers": {
    "layer_from_source1": {
      "tilesize": 256,
      "resampling": "nearest",
      "buffer": 1,
      "source_rules": {
        "filter": [
          "all",
          ["in", ["bandindex"], ["literal", [1]]],
          ["in", ["sourcetag"], ["literal", ["source1"]]]
        ]
      }
    },
    "layer_from_source2": {
      "tilesize": 256,
      "resampling": "nearest",
      "buffer": 1,
      "source_rules": {
        "filter": [
          "all",
          ["in", ["bandindex"], ["literal", [1]]],
          ["in", ["sourcetag"], ["literal", ["source2"]]]
        ]
      }
    }
  }
}
```