- INDEX
- DATA
- SYNC
- HUMAN
Short answer
How can I add new data to my LanceDB vector index using AI?
Your agent uses the LanceDB MCP to push new rows and embeddings directly into your existing tables. This triggers an immediate update to your ANN index. You get a live, updated knowledge base.
The outcomes.
Where your data lands.
What happens once your agent processes the new information.
INDEX
Immediate ANN updates
The agent pushes new embeddings into the table. Your vector search results reflect this new data immediately.
DATA
Structured storage
New rows land in your LanceDB tables alongside their vectors. Your metadata stays synced with the embeddings.
SYNC
Real-time availability
The AI ensures the new information is searchable. You don't have to wait for a batch job to finish.
HUMAN
Manual review needed
The agent inserts the data, but you still need to verify the source information is correct.
The workflow
What your AI does when data arrives.
The agent handles the database writes so you don't have to.
Data preparation
Your agent takes the raw text or structured info you provide and prepares it for the database.
insert_rowsRow insertion
The agent writes the new records and their corresponding embeddings into your LanceDB table.
insert_rowsIndex refresh
The database automatically updates the ANN index to include the newly inserted vectors.
insert_rowsVerification
The agent confirms the write was successful so you can immediately query the new data.
insert_rows
Try it
Copy these to start.
Use these prompts to trigger an update.
Starting points
These are opening lines. Swap the bracketed text for your actual table names.
LanceDB Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_85lSO635ajE8lXHCZKWQ7UC6VHfybBdy20nU97wm/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 — LanceDB capabilities are ready to use.
{
"mcpServers": {
"lancedb-serverless-vector-db-mcp": {
"url": "https://edge.vinkius.com/vk_preview_85lSO635ajE8lXHCZKWQ7UC6VHfybBdy20nU97wm/mcp"
}
}
}Add this customer feedback to my 'user_logs' table in LanceDB.
Insert these three product descriptions and their embeddings into the 'catalog' table.
Update my LanceDB index with the new documentation I just uploaded.
Push this JSON object into the 'sensor_data' table.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect LanceDB once, then ask.
One click connects your client to the MCP. Your credentials stay encrypted, and the full walkthrough lives on the Connector page.
Connect LanceDB to your AIFAQ
How this task behaves.
- 01
Does the AI automatically create new tables?
No. The agent uses the existing tables you have already set up in LanceDB.
- 02
Can the AI delete old vectors to save space?
No. This MCP only handles inserting new rows; it does not have tools for deleting or dropping data.
- 03
How fast is the index update?
The update happens as part of the insertion process, making the new vectors available for search immediately.
- 04
Do I need to provide the embeddings myself?
Your AI client handles the embedding generation before calling the tool to insert the rows.
- 05
Can the AI change the schema of my table?
No. The agent can only add rows that match your existing table structure.
More questions about LanceDB? See everything the LanceDB Connector can do
Connect LanceDB to Claude, Cursor, ChatGPT & more
