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

# Ranking

> Customize how search results are ranked and scored

Agentset provides built-in reranking, but you can also implement custom ranking logic to tune results for your specific use case.

## Built-in reranking

Agentset reranks results by default using a cross-encoder model. Configure this behavior in your search request.

<CodeGroup>
  ```typescript TypeScript theme={null}
  const results = await ns.search("product features", {
    topK: 50,
    rerank: true,
    rerankLimit: 10,
    rerankModel: "cohere:rerank-v3.5",
  });
  ```

  ```python Python theme={null}
  results = client.search.execute(
      query="product features",
      top_k=50,
      rerank=True,
      rerank_limit=10,
      rerank_model="cohere:rerank-v3.5",
  )
  ```
</CodeGroup>

See [Search](/search-and-retrieval/search#reranking) for available rerank models and parameters.

## Custom score weighting

Apply weights to boost or penalize results based on metadata. This is useful when certain document types should rank higher or lower for your use case.

<CodeGroup>
  ```typescript TypeScript theme={null}
  const results = await ns.search(query, { topK: 20 });

  // Apply weights based on document type
  const weighted = results
    .map((result) => {
      let weight = 1.0;

      // Boost official documentation
      if (result.metadata?.category === "official-docs") {
        weight = 1.2;
      }

      // Penalize older content
      if (result.metadata?.year < 2023) {
        weight = 0.8;
      }

      return {
        ...result,
        weightedScore: result.score * weight,
      };
    })
    .sort((a, b) => b.weightedScore - a.weightedScore);
  ```

  ```python Python theme={null}
  results = client.search.execute(query=query, top_k=20)

  # Apply weights based on document type
  weighted = []
  for result in results.data:
      weight = 1.0

      # Boost official documentation
      if result.metadata.get("category") == "official-docs":
          weight = 1.2

      # Penalize older content
      if result.metadata.get("year", 2024) < 2023:
          weight = 0.8

      weighted.append({
          **result.__dict__,
          "weighted_score": result.score * weight,
      })

  weighted.sort(key=lambda x: x["weighted_score"], reverse=True)
  ```
</CodeGroup>

## Combining strategies

Combine reranking with custom weighting for fine-grained control.

<CodeGroup>
  ```typescript TypeScript theme={null}
  // 1. Get reranked results
  const results = await ns.search(query, {
    topK: 50,
    rerank: true,
    rerankLimit: 20,
  });

  // 2. Apply custom weights
  const final = results
    .map((result) => {
      let score = result.score;

      // Apply domain-specific weighting
      if (result.metadata?.docType === "faq") {
        score *= 1.1; // Boost FAQs
      }

      return { ...result, finalScore: score };
    })
    .sort((a, b) => b.finalScore - a.finalScore)
    .slice(0, 10);
  ```

  ```python Python theme={null}
  # 1. Get reranked results
  results = client.search.execute(
      query=query,
      top_k=50,
      rerank=True,
      rerank_limit=20,
  )

  # 2. Apply custom weights
  final = []
  for result in results.data:
      score = result.score

      # Apply domain-specific weighting
      if result.metadata.get("doc_type") == "faq":
          score *= 1.1  # Boost FAQs

      final.append({**result.__dict__, "final_score": score})

  final.sort(key=lambda x: x["final_score"], reverse=True)
  final = final[:10]
  ```
</CodeGroup>

## Next steps

* [API Reference](/api-reference/endpoint/search) — Reranking parameters and options
* [Filtering](/search-and-retrieval/filtering) — Narrow results before ranking
* [Citations](/search-and-retrieval/citations) — Add source attribution to responses
* [AI SDK Integration](/search-and-retrieval/ai-sdk-integration) — Use Agentset with Vercel AI SDK