Vector Database Diagnostics
Run the same semantic query against the native vector stores and logical indexes in this deployment. The MCP server/tool retrieval index is exposed separately even though it is physically stored inside the shared project vector database.
| Query | Database | Kind | Collections | State | Path |
|---|---|---|---|---|---|
Shared Project Vector Storeshared |
shared | 4 | available | /persist/agent-runtime/context/vectors.sqlite3 | |
MCP Server / Tool Retrieval Indexmcp-server-info |
MCP semantic index | 1 | available | /persist/agent-runtime/context/vectors.sqlite3 | |
Artifact Semantic Indexartifact-index |
artifact semantic index | 0 | available Artifact semantic index has not been initialized yet (no artifact_* collection). |
/persist/agent-runtime/context/vectors.sqlite3 | |
Agent memory: Lumen (agent-lumen_fast_uncensored-1af60847)memory:agent-lumen_fast_uncensored-1af60847 |
memory | 0 | available | /persist/agent-runtime/context/memory_stores/agent-lumen_fast_uncensored-1af60847/vectors/vectors.sqlite3 | |
Agent memory: Uncensored (agent-uncensored-agent-f7aae8b4)memory:agent-uncensored-agent-f7aae8b4 |
memory | 0 | available | /persist/agent-runtime/context/memory_stores/agent-uncensored-agent-f7aae8b4/vectors/vectors.sqlite3 | |
Lumen Mind (lumen)memory:lumen |
memory | 1 | available | /persist/agent-runtime/context/memory_stores/lumen/vectors/vectors.sqlite3 |
Query
One embedding is generated for the query, then the selected databases are searched with identical vector input. Serial runs await each database in order; parallel runs query the selected databases concurrently.