mcp-memoria
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MEMORIA_DB | No | Path to the SQLite database file. | |
| MEMORIA_CONFIG | No | Path to the memoria.config.json configuration file. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| buscar_memoriaA | Busca en la memoria del usuario (notas, memorias de agentes y READMEs indexados) y devuelve fragmentos con cita exacta |
| leer_fragmentoA | Lee líneas exactas de un archivo ya indexado. Sirve para ampliar una cita devuelta por buscar_memoria sin que el modelo tenga que leer el archivo entero. |
| guardar_notaA | Guarda un dato nuevo en la memoria (una decisión, una preferencia, un dato del usuario) y lo deja indexado. Si el texto ya existía no lo duplica: devuelve la cita original. Usala cuando el usuario diga algo que convenga recordar en próximas sesiones. |
| listar_fuentesA | Lista las fuentes indexadas (archivos, etiqueta, tamaño y fecha de indexado) y el resumen del índice. Sirve para saber qué memoria está disponible antes de buscar. |
| reindexarA | Vuelve a recorrer las fuentes configuradas y actualiza el índice (solo procesa archivos cambiados). Devuelve cuántos archivos son nuevos, cuántos se actualizaron y cuántos se borraron. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 5 tools
Each tool targets a distinct stage of the memory lifecycle: listing sources, searching, expanding a citation, writing a new note, and rebuilding the index. There is no plausible confusion between them, and the descriptions explicitly chain them (buscar_memoria → leer_fragmento, listar_fuentes before searching).
Four of five names follow a clear Spanish verb_noun pattern (listar_fuentes, buscar_memoria, leer_fragmento, guardar_nota). reindexar is a bare verb without a noun, a minor deviation but still readable and unambiguous.
Five tools is a well-scoped surface for a personal memory/index server; each maps to a real operation with no filler. Nothing feels thin because the read/write/search/reindex loop is fully covered.
The core loop (inspect index, search, read, write, reindex) is covered, including deduplication on write and incremental reindexing. There is no explicit way to delete or edit an existing note, so stale memories can only be removed via file changes plus reindexing.