- PRECISION
- ACCURACY
- EFFICIENCY
- HUMAN
Short answer
How do I rerank search results with my AI?
Use the rerank tool to re-sort your retrieved documents based on actual semantic relevance. This ensures your agent processes the most important information first. You'll get much cleaner context for your prompts.
The results.
Where your context lands.
The reranking process changes how your agent sees your data.
PRECISION
Top-tier relevance
The rerank tool moves the most useful documents to the front of the list. Your agent sees exactly what it needs immediately.
ACCURACY
Better answers
By filtering out noise, the AI avoids hallucinating based on irrelevant snippets. You get responses that actually match the user's intent.
EFFICIENCY
Reduced noise
The agent ignores the junk at the bottom of the search results. This keeps the context window focused on high-signal data.
HUMAN
Manual review
You decide if the reranked list is sufficient. The AI doesn't delete any original data, so you can always check the full set.
The workflow
What your AI does when search hits arrive.
Instead of just grabbing the first few matches, your agent uses a second pass to verify importance.
Fetch initial hits
Your agent pulls a broad set of potential matches from your vector database or search index.
rerankApply reranking
The rerank tool evaluates the relationship between the query and each document to find the true matches.
rerankReorder context
The most relevant snippets move to the top of the stack, pushing irrelevant data down.
rerankFeed the prompt
The agent takes this newly ordered list and injects it into the prompt for a final response.
rerank
Try it
Copy these to start.
Use these prompts to begin reranking your data.
Starting points
These are starting points. Swap out the specific document names or folder paths for your own.
Voyage AI Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_x7lwFbbAdz5StvxoOu2Tkh9eDMu9Uh5i3GKObLMP/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Voyage AI capabilities are ready to use.
{
"mcpServers": {
"voyage-ai-ai-embeddings-api-mcp": {
"url": "https://edge.vinkius.com/vk_preview_x7lwFbbAdz5StvxoOu2Tkh9eDMu9Uh5i3GKObLMP/mcp"
}
}
}Rerank the results from my recent PDF search in the documentation folder.
Take the top 20 hits from my customer support logs and rerank them by relevance to this query.
Use the rerank tool to sort the retrieved chunks from the project wiki.
Rerank the search results from my GitHub repo snippets to find the specific function definition.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Voyage AI once, then ask.
Just link your account once. Your credentials stay encrypted, and you can start reranking immediately in your preferred client.
Connect Voyage AI to your AIFAQ
How reranking behaves
- 01
Does reranking change my original database?
No. The rerank tool only changes the order of the results sent to your agent. Your underlying data stays exactly as it was.
- 02
Can the AI delete documents that aren't relevant?
No. The tool only reorders the list. It doesn't delete or modify any of your source files.
- 03
Will this work with any vector database?
Yes, as long as your agent can retrieve the initial list of documents to pass to the rerank tool.
- 04
Do I need to change my embedding model?
No. You can use your existing embeddings for the initial search and use Voyage AI specifically for the reranking step.
- 05
How many documents can I rerank at once?
The tool handles the list of documents your agent retrieves, allowing you to sort through dozens or hundreds of hits.
More questions about Voyage AI? The Connector page answers them. See everything the Voyage AI Connector can do
Connect Voyage AI to Claude, Cursor, ChatGPT & more
