- SYNC
- CLEAN
- HUMAN
- READY
Short answer
How can I update my Marqo search index using an AI agent?
Your agent uses specific tools to push new JSON documents into your indices or wipe out outdated records. This keeps your vector search data fresh and accurate. You get a perfectly synced index without manual database management.
Index management outcomes
Where your data lands
The agent moves data in and out of your Marqo indices based on your instructions.
SYNC
Fresh search data
The agent adds new JSON documents to your index. Your vector search immediately includes this new information.
CLEAN
Removed stale records
The agent deletes specific documents from the index. This prevents old or incorrect data from appearing in search results.
HUMAN
Manual oversight
You decide which specific records are outdated. The agent waits for your command before it touches your index.
READY
Accurate retrieval
The index stays updated with your latest data. Your search queries return the most current information available.
The workflow
What your AI does when <the thing arrives>.
The agent acts as the bridge between your instructions and the Marqo database.
Analyze request
Your agent looks at your new data or your request to remove entries. It determines which specific Marqo action is required.
add_documentsdelete_documentsInsert documents
The agent pushes new JSON documents into the vector store. This expands your searchable dataset.
add_documentsRemove records
The agent wipes specific documents from the index. This keeps the search space clean.
delete_documentsFinalize index
The agent completes the requested changes. Your index is now updated and ready for the next query.
add_documentsdelete_documents
Try it
Copy these to start.
Use these prompts to trigger index updates.
Starting points
These are opening lines. Swap the bracketed text for your actual JSON data or document IDs.
Marqo AI Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_yWwQAD75WfnpbXY72FMh8SBRNTeZWrryXTJG2Ifa/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 — Marqo AI capabilities are ready to use.
{
"mcpServers": {
"marqo-ai-vector-search-embeddings-mcp": {
"url": "https://edge.vinkius.com/vk_preview_yWwQAD75WfnpbXY72FMh8SBRNTeZWrryXTJG2Ifa/mcp"
}
}
}Add this JSON document to my Marqo index: [insert JSON here]
Delete the document with ID [insert ID] from my Marqo index.
I have new product data in this list, please add it to Marqo: [insert data]
Remove all outdated records from the index using these IDs: [insert IDs]
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Marqo AI once, then ask.
Connect your Marqo account in one click. Your credentials stay encrypted, and you can immediately start managing your indices through your agent.
Connect Marqo AI to your AIFAQ
How this task behaves
- 01
Can the AI create a brand new index from scratch?
No. The agent can only add or delete documents within existing indices.
- 02
Will the AI delete my entire database?
No. The agent only deletes specific documents that you identify.
- 03
Can I upload a whole CSV file at once?
No. The agent works with JSON documents. You should provide the data in JSON format for the agent to process.
- 04
Does the AI decide which documents are outdated?
No. You must tell the agent which specific documents or IDs need to be removed.
- 05
Can the AI change the embedding model used in the index?
No. The agent only manages the documents within the current index configuration.
More questions about Marqo AI? The Connector page answers them. See everything the Marqo AI Connector can do
Connect Marqo AI to Claude, Cursor, ChatGPT & more
