delete_memory
Delete memory by ID.
LABSCO SUMMARY
It keeps an agent's notes and your codebase in one local store, and the retrieval side is the reason to reach for it.
recall is the single entry point worth learning: it blends semantic, keyword and graph results in one call instead of making you choose. Memories are versioned rather than overwritten — invalidate soft-deletes and points at the replacement, and get_valid replays what was true at any timestamp you name.
What it is
A memory server written in Rust that ships as a single self-contained executable: SurrealDB, the embedding model and the graph all run inside the process, so there is no separate vector store or graph database to stand up.
What you get
- Store, update, list and delete memories, and pull any one of them back in full by ID
- recall as the default way in — it fuses vector, BM25 and graph hits through reciprocal rank fusion so you do not have to pick a search mode
- search_memory when you do want to pick: vector for semantic similarity, bm25 for exact keyword matching
- Point-in-time reads — get_valid takes an ISO 8601 timestamp and returns the memories that were valid then
- invalidate soft-deletes a memory with a reason and can link the memory that supersedes it
- Code indexing: index a directory once, then search it by meaning, by symbol name, or by walking the call graph from a symbol
- A knowledge graph you can write to — create entities and typed relations, fetch what is related at a given depth and direction, and detect communities
Requirements
No account and no API key — the embedding model is baked into the binary and runs on CPU. The Docker path wants three things on the run command: a named volume at -v mcp-data:/data so the database and the ~1.2 GB model survive restarts, your code at -v $(pwd):/project:ro for indexing, and --init for a clean shutdown. Also published to npm as memory-mcp-1file.
Setup effort
One command — claude mcp add memory -- npx -y memory-mcp-1file
USE CASES
01
Picking a task back up after the context window resets
Search the stored memories for the task that was in progress, then pull the full note back by ID.
02
Asking one question without choosing a search mode
recall fuses vector, keyword and graph results in a single call and ranks them together.
03
Making an unfamiliar codebase searchable
Index the directory once, then look things up by meaning or by symbol name.
04
Tracing what calls a function before you change it
Find the symbol, then walk its callers and callees to see the blast radius.
05
Retiring a decision without losing the trail
Soft-delete the old memory, link the one that replaces it, then replay what was valid on an earlier date.
06
Mapping how people, systems and decisions relate
Create entities and typed relations, then pull everything related to one of them at a chosen depth.
19 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
delete_project
Delete indexed project.
get_memory
Get full memory by ID.
get_status
Get system status and startup progress.
get_valid
Get valid memories. Optional timestamp (ISO 8601) for point-in-time query.
how_to_use
Show all available tools with usage examples and parameter combinations.
index_project
Index codebase directory for code search.
invalidate
Soft-delete memory, optionally linking replacement.
knowledge_graph
Knowledge graph ops. Actions: create_entity(name, entity_type?, description?) | create_relation(from_entity, to_entity, relation_type, weight?) | get_related(entity_id, depth?, direction?) | detect_communities()
list_memories
List memories (newest first).
project_info
Project info. Actions: list() | status(project_id) | stats(project_id)
recall
Best memory retrieval (query). Combines vector+BM25+graph via RRF fusion. Use as default.
recall_code
Code retrieval (query, mode?: vector|hybrid). Default hybrid = vector+BM25+graph fusion. Filters: path_prefix?, language?, chunk_type?
reset_all_memory
DANGER: Reset all database data (requires confirm=true).
search_memory
Search memories (query, mode?: vector|bm25). Vector=semantic similarity, bm25=exact keyword match.
search_symbols
Search code symbols by name.
store_memory
Store a new memory.
symbol_graph
Navigate symbol call graph. Actions: callers(symbol_id) | callees(symbol_id) | related(symbol_id, depth?, direction?)
update_memory
Update memory fields.
BEFORE YOU INSTALL
What to know first
Tools you get
19 tools when we connected.
Requires
npx on PATH
Where it runs
All 8 documented setups run the server on your machine.
CONTEXT COST
Light Standing cost
Rated by the size of the 19 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
Claude Code
TERMINAL
claude mcp add memory -- npx -y memory-mcp-1fileMCP Clients
Claude Desktop — Docker
CLAUDE DESKTOP — DOCKER · JSON
{
"mcpServers": {
"memory": {
"command": "docker",
"args": [
"run",
"--init",
"-i",
"--rm",
"--memory=3g",
"-v",
"mcp-data:/data",
"-v",
"<YOUR_PATH_TO_PROJECT>:/project:ro",
"ghcr.io/pomazanbohdan/memory-mcp-1file:latest"
]
}
}
}Claude Desktop — npx, Cursor — npx, Windsurf — npx, Gemini CLI — npx
JSON
{
"mcpServers": {
"memory": {
"command": "npx",
"args": [
"-y",
"memory-mcp-1file"
]
}
}
}VS Code — npx
VS CODE — NPX · JSON
{
"servers": {
"memory": {
"command": "npx",
"args": [
"-y",
"memory-mcp-1file"
]
}
}
}Any MCP client — bunx
ANY MCP CLIENT — BUNX · JSON
{
"mcpServers": {
"memory": {
"command": "bunx",
"args": [
"memory-mcp-1file"
]
}
}
}
