Vector-only similarity search (semantic search). Uses vector embeddings for semantic similarity matching. Requires embeddings to be generated (use memvid_put with embed=true). Args: file: Path to the .mv2 memory file query: Search query text (will be embedded) limit: Maximum results (default: 10)...
AI agents call memvid_vec_search to retrieve information from Memvid without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves and queries data from memory files using vector embeddings but produces no side effects—it neither creates, modifies, deletes, executes code, nor commits financial transactions. It is a pure read operation analogous to a search or query function.
From the tool's definition Tool performs 'Vector-only similarity search (semantic search)' and 'Returns: Array of results with cosine similarity scores'. The description explicitly states it conducts similarity matching and retrieval without modifying data.
Attacks that exploit this kind of access
Vector-only similarity search (semantic search). Uses vector embeddings for semantic similarity matching. Requires embeddings to be generated (use memvid_put with embed=true). Args: file: Path to the .mv2 memory file query: Search query text (will be embedded) limit: Maximum results (default: 10) Returns: Array of results with cosine similarity scores. It is categorised as a Read tool in the Memvid MCP Server, which means it retrieves data without modifying state.
Register the Memvid MCP server in PolicyLayer and add a rule for memvid_vec_search: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Memvid. Nothing to install.
memvid_vec_search is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the memvid_vec_search rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for memvid_vec_search. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
memvid_vec_search is provided by the Memvid MCP server (tapiocapioca/memvid-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
memvid_vec_search is one line of Memvid's registry record.
The record carries the whole server: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.
Teams ship this data inside their own products. See what a licence covers →