Ahnlich MCP
Ahnlich MCP gives MCP-compatible agents access to Ahnlich vector storage, semantic search, and metadata filtering.
It runs locally over stdio and does not require cloud credentials.
Profiles
| Profile | Input | Required services |
|---|---|---|
ai | Raw text | Ahnlich AI and Ahnlich DB |
db | Precomputed embeddings | Ahnlich DB |
Use ai when the agent should work with text. Ahnlich AI generates the
embeddings before storing them in Ahnlich DB.
Use db when your application already generates embeddings.
Exposed tools
Both profiles expose the same 11 tool names.
| Tool | Purpose | Access |
|---|---|---|
ping | Check whether the configured Ahnlich service is reachable | Read |
server_info | Inspect the service version, address, and memory information | Read |
create_store | Create a text or embedding store | Write |
list_stores | List existing stores | Read |
drop_store | Delete a store and all its entries | Destructive |
store_entries | Store text or precomputed embeddings with metadata | Write |
similarity_search | Find entries that are semantically or numerically similar | Read |
get_by_metadata | Retrieve entries matching indexed metadata | Read |
delete_by_metadata | Delete entries matching indexed metadata | Destructive |
create_predicate_index | Index metadata keys for filtering | Write |
drop_predicate_index | Remove metadata indexes | Destructive |
The inputs for create_store, store_entries, and similarity_search depend
on the active profile.
See the tool reference for complete inputs, return types, examples, and agent guidance.
Current scope
The AI profile currently accepts text. Direct image and audio tools are not yet available through Ahnlich MCP.
Ahnlich MCP connects to existing Ahnlich services; it does not start them automatically.