> ## Documentation Index
> Fetch the complete documentation index at: https://www.octogen.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# More Like This strategy recipes

> Five ready-to-use recommendation strategies for the More Like This endpoint, with real requests and live results: same-look, attribute matching, broad discovery, hybrid retrieval/ranking, and price-bounded similarity.

The same `POST /v1/products/more-like-this` endpoint powers distinct recommendation surfaces depending on which embedding columns drive retrieval and ranking. This guide shows five strategies with real requests and live results from indexed catalogs. Parameter reference: [More Like This API](/docs/api-reference/more-like-this). A visual version of this guide — the same strategies with product imagery — lives at the [interactive demo](https://www.octogen.ai/demos/mlt-api.html).

The examples below use two source products:

* **Kumi Cropped Jacket** — apparel, catalog `lagence_merchant`, referenced by UUID (`source.uuid`). Enriched styles: classic, sophisticated, military-inspired, polished, preppy. Tags: cropped, 3/4 sleeves, gold buttons, embossed buttons, shoulder pads, flap pockets.
* **Epi Mabillon Backpack Tassil** — luxury resale, catalog `fashionphile`, referenced by URL (`source.url`). Enriched styles: luxury, minimalist. Tags: textured leather, silver-tone hardware, adjustable straps, zip closure.

Every request works with either source form. `debug: true` is included throughout so the response's `effectiveQuery` shows the exact query the server executed — remove it in production.

## Choosing a strategy

| You want                                  | Strategy                          | Key parameters                                              |
| ----------------------------------------- | --------------------------------- | ----------------------------------------------------------- |
| "Same look" recommendations               | Same look (default)               | none                                                        |
| Products sharing concrete attributes      | Closest attribute matches         | `retrieval_embedding_columns: ["attributes_embedding"]`     |
| Wider discovery                           | Broad semantic matches            | `retrieval_embedding_columns: ["embedding"]`                |
| Max attribute matches + closeness of look | Attribute retrieval, style-ranked | retrieval `attributes_embedding`, ranking `style_embedding` |
| Cheaper (or premium) alternatives         | Price-bounded same look           | `price_preference: "lower"` / `"higher"`                    |

Facet matching is automatic in every strategy: when you don't pass `include_facets` or `exclude_facets`, the server derives audience facets from the source product's enrichment. Pass your own facets (for example `color_family`) on top of any strategy when you want explicit control.

## Same look (default)

Omit the embedding-column parameters and the server picks the strategy: query text from the source product's enriched styles and tags, retrieval over the style and tags embeddings, plus automatic audience facets (product type, gender, age group).

```json Request theme={null}
{
  "source": {
    "uuid": "19daa0d1-8d4a-4cb8-aaa6-11a98b005e10"
  },
  "catalog": "lagence_merchant",
  "limit": 6,
  "debug": true
}
```

Top results (live data):

```json theme={null}
[
  {
    "title": "Kumi Cropped Jacket",
    "currentPrice": 575.0
  },
  {
    "title": "Kumi Cropped Jacket",
    "currentPrice": 525.0
  },
  {
    "title": "Jia Cropped Tweed Jacket",
    "currentPrice": 397.5
  }
]
```

`effectiveQuery` echoes `retrievalEmbeddingColumns: ["style_embedding", "tags_embedding"]`.

## Closest attribute matches

Retrieve via the attributes embedding so candidates share concrete product attributes — construction, hardware, closures, silhouette — rather than overall styling.

```json Request theme={null}
{
  "source": {
    "url": "https://www.fashionphile.com/products/louis-vuitton-epi-mabillon-backpack-tassil-64262"
  },
  "catalog": "fashionphile",
  "limit": 6,
  "retrieval_embedding_columns": [
    "attributes_embedding"
  ],
  "debug": true
}
```

Top results (live data):

```json theme={null}
[
  {
    "title": "Epi Mabillon Backpack",
    "currentPrice": 675.0,
    "productUrl": "https://www.fashionphile.com/products/louis-vuitton-epi-mabillon-backpack-moka-303843"
  },
  {
    "title": "Metallic Calfskin Marti Backpack Bag Light Blue Rhodium Hardware",
    "currentPrice": 265.0,
    "productUrl": "https://www.fashionphile.com/products/alexander-wang-metallic-calfskin-marti-backpack-bag-light-blue-rhodium-hardware-1708668"
  },
  {
    "title": "Calfskin Goya Backpack",
    "currentPrice": 995.0,
    "productUrl": "https://www.fashionphile.com/products/loewe-calfskin-goya-backpack-scarlet-1847819"
  }
]
```

`effectiveQuery` echoes `retrievalEmbeddingColumns: ["attributes_embedding"]`.

## Broad semantic matches

Retrieve via the general-purpose base embedding for wider, discovery-style results.

```json Request theme={null}
{
  "source": {
    "uuid": "19daa0d1-8d4a-4cb8-aaa6-11a98b005e10"
  },
  "catalog": "lagence_merchant",
  "limit": 6,
  "retrieval_embedding_columns": [
    "embedding"
  ],
  "debug": true
}
```

Top results (live data):

```json theme={null}
[
  {
    "title": "Kumi Cropped Jacket",
    "currentPrice": 575.0
  },
  {
    "title": "Kumi Cropped Jacket",
    "currentPrice": 525.0
  },
  {
    "title": "Jia Tweed Jacket",
    "currentPrice": 437.5
  }
]
```

`effectiveQuery` echoes `retrievalEmbeddingColumns: ["embedding"]`.

## Max attribute matches, ordered by look

Two stages working together: retrieve candidates by attribute closeness, then re-score them by style-embedding similarity.

This is the combination for related-products rails that should respect concrete attributes but present the closest looks first.

```json Request theme={null}
{
  "source": {
    "url": "https://www.fashionphile.com/products/louis-vuitton-epi-mabillon-backpack-tassil-64262"
  },
  "catalog": "fashionphile",
  "limit": 6,
  "retrieval_embedding_columns": [
    "attributes_embedding"
  ],
  "ranking_embedding_columns": [
    "style_embedding"
  ],
  "debug": true
}
```

Top results (live data):

```json theme={null}
[
  {
    "title": "Epi Mabillon Backpack",
    "currentPrice": 675.0,
    "productUrl": "https://www.fashionphile.com/products/louis-vuitton-epi-mabillon-backpack-moka-303843"
  },
  {
    "title": "Calfskin Goya Backpack",
    "currentPrice": 995.0,
    "productUrl": "https://www.fashionphile.com/products/loewe-calfskin-goya-backpack-scarlet-1847819"
  },
  {
    "title": "Epi Mabillon Backpack Black",
    "currentPrice": 850.0,
    "productUrl": "https://www.fashionphile.com/products/louis-vuitton-epi-mabillon-backpack-black-254229"
  }
]
```

`effectiveQuery` echoes `retrievalEmbeddingColumns: ["attributes_embedding"]` and `rankingEmbeddingColumns: ["style_embedding"]`.

## Same look, lower price

The default strategy constrained to products priced below the source. `"higher"` works the same way in the other direction.

```json Request theme={null}
{
  "source": {
    "uuid": "19daa0d1-8d4a-4cb8-aaa6-11a98b005e10"
  },
  "catalog": "lagence_merchant",
  "limit": 6,
  "price_preference": "lower",
  "debug": true
}
```

Top results (live data):

```json theme={null}
[
  {
    "title": "Janu Cropped Knit Jacket",
    "currentPrice": 297.5
  },
  {
    "title": "Jia Cropped Tweed Jacket",
    "currentPrice": 287.5
  },
  {
    "title": "Jia Tweed Jacket",
    "currentPrice": 172.5
  }
]
```

`effectiveQuery` echoes `retrievalEmbeddingColumns: ["style_embedding", "tags_embedding"]`.

## Combining with facets

Any strategy accepts explicit facets. For example, attribute-focused retrieval pinned to the source's color family:

```json Request theme={null}
{
  "source": {
    "url": "https://www.fashionphile.com/products/louis-vuitton-epi-mabillon-backpack-tassil-64262"
  },
  "catalog": "fashionphile",
  "retrieval_embedding_columns": [
    "attributes_embedding"
  ],
  "include_facets": [
    {
      "name": "color_family",
      "values": [
        "black"
      ]
    }
  ],
  "limit": 12
}
```

See the [More Like This API reference](/docs/api-reference/more-like-this) for all parameters, accepted embedding-column values, and error semantics.