AI & Machine Learning

longmem.dev

Provides semantic search and management of long-term memory for AI agents or personal knowledge.

1 endpoint10 known toolsFirst detected June 16, 2026Last detected August 26, 2026

ENDPOINT 1

https://longmem.dev/v1/mcp

No auth detected

MCP server metadata

Name: longmem
Version: 0.1.0

Capabilities

tools

Server instructions

LongMem — shared long-term memory for agents. Use search_memory to recall, store_memory to remember (omit collection in search to search everything). To discover patterns another agent recorded: related_memory finds neighbours, what_changed shows superseded facts, collection_health flags drift, summarize_collection gives a board-report briefing of the whole collection. To keep shared memory healthy: find_duplicates surfaces redundancy, archive_stale retires old memories (reversible; dry_run by default).

Known tools 10

search_memory

Semantic search over the caller's long-term memory.

Inferred read-only

get_memory

Fetch the full text of a single memory by id, plus a download URL for its file if any.

Inferred read-only

store_memory

Store a new memory (a fact, note, or document text) for later recall.

Inferred read-only

list_collections

List the caller's memory collections with item counts.

Inferred read-only

related_memory

Find the memories most semantically related to a given memory id — its nearest neighbours in the same collection.

Inferred read-only

what_changed

Show where the current truth in a collection superseded an older fact (graph 'updates' edges) — each {current, superseded} pair, newest first.

Inferred read-only

collection_health

At-a-glance health of a shared collection as agents write to it: total / current / stale / expired / supersessions and a single needs_attention roll-up.

Potential side effects

summarize_collection

Roll a shared collection into a board-report-style briefing: a concise summary plus key themes, drawn from the most important recent memories.

Inferred read-only

find_duplicates

Surface near-duplicate memories in a collection (cosine >= threshold) as clusters — read-only, deletes nothing.

Inferred read-only

archive_stale

Soft-archive memories not updated in `older_than_days` days (skips pinned and already-archived).

Inferred read-only

CONNECT WITH APPROVAL

Client installation

Review this server and its permissions before adding it. Secret placeholders must be set locally.

Codex

~/.codex/config.toml

[mcp_servers.longmem]
url = "https://longmem.dev/v1/mcp"
enabled = true

Claude Code

.mcp.json

{
  "mcpServers": {
    "longmem": {
      "type": "http",
      "url": "https://longmem.dev/v1/mcp"
    }
  }
}

Claude Desktop

Settings → Connectors → Add custom connector

Name: longmem
Remote MCP URL: https://longmem.dev/v1/mcp

Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.

Cursor

.cursor/mcp.json

{
  "mcpServers": {
    "longmem": {
      "url": "https://longmem.dev/v1/mcp"
    }
  }
}

Visual Studio Code

.vscode/mcp.json

Add to Visual Studio Code

{
  "servers": {
    "longmem": {
      "type": "http",
      "url": "https://longmem.dev/v1/mcp"
    }
  }
}

Generic MCP

Client-specific MCP configuration

{
  "name": "longmem",
  "transport": "streamable-http",
  "url": "https://longmem.dev/v1/mcp"
}

MCP Inspector

Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.

TRUST AND VERIFICATION EVIDENCE

Trust Data Available

BuiltWith Trust API v2 evidence for longmem.dev was fetched 2026-08-24T04:08:30.914Z and is being refreshed.

Trust status Neutral

longmem.dev is assessed as Neutral: No suspicious signals found, but no strong positive signal either

Indexed

Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.