Rememb
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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 | {
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| rememb_list_storesA | List configured memory stores for this MCP process: global (~/.rememb) plus any project stores passed via --project. Safe, read-only. |
| rememb_getA | Fetch one memory entry by ID with full content. Safe, read-only. Use after rememb_search or rememb_recent when you already know the entry ID. |
| rememb_recentA | Return recently updated entries, newest first. Safe, read-only. Useful for catching up on what changed since the last session. |
| rememb_list_tagsA | List tags used in the store with usage counts. Safe, read-only. Use to discover filters before rememb_search or rememb_read_page. |
| rememb_readA | Read all memory entries or filter by section. Safe, read-only operation with no side effects. Returns full entry content; summarize task-relevant facts in your working context after large reads. |
| rememb_read_pageB | Read a paginated slice of entries with optional section or tag filtering. Returns full entry content; summarize task-relevant facts in your working context after each page. |
| rememb_searchA | Keyword and token search over memory entries. Safe, read-only. Returns id, score, tags, date, and a short excerpt. Use rememb_get for full body. |
| rememb_versionsB | List all revisions known for an entry, including prior content snapshots and deleted states when requested. Safe, read-only operation. |
| rememb_restoreA | Restore a soft-deleted entry, or restore a specific prior version as the new current head. Restores are non-destructive: they append a new head revision rather than rewriting history. |
| rememb_diffA | Show a unified diff between two revisions of the same entry. Safe, read-only operation. |
| rememb_writeA | Save a new memory entry or multiple entries in one call. Existing entries are never overwritten. Use rememb_edit to update an existing entry by ID. |
| rememb_editA | Update one existing memory entry by ID or multiple entries in one call via updates[]. Modifies only the fields provided (content, section, or tags) — omitted fields are unchanged. Non-destructive: entries are updated, not deleted and recreated. Use rememb_write to create new entries, rememb_delete to permanently remove them. |
| rememb_deleteA | Soft-delete one memory entry by its ID or multiple entries via entry_ids[]. Deleted entries are hidden by default from reads and search, but can be restored later. Use rememb_restore to undo a deletion. Use rememb_clear to permanently delete all entries at once. |
| rememb_clearA | Permanently delete ALL memory entries at once. Irreversible — no recovery is possible after this operation. Requires confirm=true as a safety guard. Use rememb_delete to remove a single entry by ID instead. Only use this to fully reset the memory store. |
| rememb_statsA | Return memory usage statistics: total entries, size in KB, oldest and newest entry dates, and count per section. Safe, read-only operation with no side effects. Use to give the user an overview of their memory store or to decide if cleanup is needed. |
| rememb_consolidateA | Consolidate literal duplicate entries and merge metadata (tags and access data). Exact normalized-content match only; review near-duplicates with the agent before consolidating. |
| rememb_list_skillsA | List bundled rememb skills discovered from the installed package contents. Safe, read-only operation. |
| rememb_use_skillA | Load one bundled rememb skill by identifier or exact declared name and return its instructions. Safe, read-only operation. Use rememb_list_skills first to inspect available skills. |
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 18 tools
Core entry operations are clearly separated: write/create, edit/update, delete/soft-delete, restore, clear, get, search, recent, and versions/diff each target distinct intents. The main overlap is between rememb_read and rememb_read_page, which both return full entry content and differ only by pagination/filtering; search/recent/get are distinguishable by description.
All tools consistently use the rememb_ prefix and mostly follow a verb or verb_noun pattern such as list_stores, get, search, write, edit, delete, restore, and read_page. A few names are noun/adjective-based (versions, recent, stats) rather than strict verbs, but the overall convention remains readable and predictable.
At 18 tools the surface is borderline heavy for a memory store, with several overlapping read/list variants (read, read_page, recent, search, get) plus auxiliary skill and store-management tools. Each tool has a plausible purpose, but some read variants could likely be consolidated to reduce selection burden.
The memory lifecycle is well covered: create, read, search, update, soft-delete, restore, permanent clear, revisions, diffing, tagging, statistics, store listing, consolidation, and skill discovery/loading. No obvious CRUD or lifecycle gaps remain for the stated domain.