recall_memoriesLABSCO SUMMARY
Storing the memory is half the job; linking it is the other half.
The rule the docs push hardest is to follow every store with a relationship - solution solves problem, fix addresses error. That is what turns the store from a pile of notes into something that answers why. The core profile is the default deliberately: the removed tools were dropped to keep the tool list from eating the context window.
What it is
A graph-backed memory store. Flat notes keep entries isolated; here memories carry typed relationships, so asking what happened with retry logic returns the causal chain rather than three unrelated notes.
What you get
- Recall as the first move - fuzzy matching with relationships included, tuned for natural-language questions
- A stricter search for exact matching, tag filters, importance thresholds and pagination when recall is too loose
- Storing typed memories - solution, problem, error, fix, decision, pattern - and reading one back in full
- Relationships between memories, with over 35 types across causation, solution, context, dependency, knowledge, comparison and workflow
- Exploring what a memory connects to, updating it, and deleting it
- Recent activity and store statistics for an overview
Requirements
uv, which runs the pinned memorygraphMCP 0.12.4 release from PyPI. The repository has since moved to a Bun command-line tool, and this release still serves MCP. Memories go into a SQLite database by default. The MCP surface ships in two profiles: a core profile of 9 tools that covers memory CRUD, relationships and discovery, and an extended profile of 12 that adds database statistics, complex queries and contextual search. Projects share one store by default - set a different path per project, or tag memories with the project name and filter on it.
Setup effort
One command — uvx --with httpx --with 'mcp<2' --from memorygraphMCP==0.12.4 memorygraph
USE CASES
01
Picking up a task the team has half-solved before
Recall against the task description first; matched memories come back with their relationships, so the earlier attempt arrives with its outcome attached.
02
Recording a fix so it is findable later
Store the problem and the solution as separate memories, then link them - without the link the pair is two orphans.
03
Finding something recall keeps missing
Switch to the stricter search with exact matching, tag filters or an importance threshold, and page through the results.
04
Correcting a memory that has gone stale
Read it first, then update it or delete it, rather than storing a contradicting entry beside it.
05
Seeing what the store has been learning
Read recent activity and the statistics to check the memory is growing where the work is.
12 tools
Listed in project docs. Your client may show a different list.
store_memorycreate_relationshipsearch_memoriesget_memoryget_related_memoriesupdate_memorydelete_memoryget_recent_activityget_memory_statisticssearch_relationships_by_contextcontextual_searchBEFORE YOU INSTALL
What to know first
Tools you get
12 tools named in the project's documentation.
Where it runs
All 5 documented clients run the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
uvx --with httpx --with 'mcp<2' --from memorygraphMCP==0.12.4 memorygraphMCP Clients
Claude Desktop, Cursor, Cline, Windsurf
JSON
{
"mcpServers": {
"memorygraph": {
"command": "uvx",
"args": [
"--with",
"httpx",
"--with",
"mcp<2",
"--from",
"memorygraphMCP==0.12.4",
"memorygraph"
]
}
}
}VS Code
VS CODE · JSON
{
"servers": {
"memorygraph": {
"command": "uvx",
"args": [
"--with",
"httpx",
"--with",
"mcp<2",
"--from",
"memorygraphMCP==0.12.4",
"memorygraph"
]
}
}
}
