Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
MEMEX_DB_PATHNoOverride the database path. Defaults to ~/Library/Application Support/memex (macOS), %LOCALAPPDATA%\memex (Windows), or ~/.local/share/memex (Linux).
MEMEX_SYNC_AUTONoSet to 'true' to enable auto-sync (default false unless setup --sync used).
ANTHROPIC_API_KEYNoAnthropic API key required for auto-summaries.
MEMEX_EXPORTS_DIRNoOverride the exports directory.
MEMEX_EMBED_BACKENDNoEmbedding backend: 'fastembed' (default) or 'ollama'.
MEMEX_SUMMARY_MODELNoModel for summaries (default 'claude-haiku-4-5-20251001').
MEMEX_REMOTE_BASE_URLNoPublic HTTPS base URL for the remote MCP connector (required for remote mode).
MEMEX_SUMMARY_ENABLEDNoSet to 'true' to enable auto-summaries (default off).
MEMEX_SYNC_ON_CAPTURENoSet to 'false' to disable sync on capture (default true).
MEMEX_EMBED_BATCH_SIZENoBatch size for embedding. Lower to reduce memory usage (default 32).
MEMEX_GITHUB_CLIENT_IDNoGitHub OAuth App Client ID (required for remote mode).
MEMEX_SUMMARY_MAX_TOKENSNoMax tokens for summaries (default 200).
MEMEX_GITHUB_CLIENT_SECRETNoGitHub OAuth App Client Secret (required for remote mode).
MEMEX_SYNC_MAX_BATCH_BYTESNoMax bytes per sync batch transfer (default 8 MB).
MEMEX_INGEST_MAX_BODY_BYTESNoMax body bytes for ingest (default from sync batch).
MEMEX_SYNC_INTERVAL_SECONDSNoInterval between automatic syncs (default 900).
MEMEX_REMOTE_ALLOWED_GITHUB_LOGINSNoComma-separated list of allowed GitHub usernames (required for remote mode).
MEMEX_SYNC_ON_CAPTURE_DEBOUNCE_SECONDSNoDebounce window in seconds for sync on capture (default 10).

Capabilities

Features and capabilities supported by this server

CapabilityDetails
tools
{
  "listChanged": true
}
logging
{}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
extensions
{
  "io.modelcontextprotocol/ui": {}
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
search_chatsA

Access to the user's persistent memory: ALL their past Claude.ai chats.

This is the only tool available to access the user's real history. Your native memory starts empty in every Claude Code session; anything the user has discussed on claude.ai lives only here.

USE PROACTIVELY (without the user asking explicitly) when:

  • They mention prior conversations or decisions: "remember when...", "we talked about...", "the other day we discussed...", "in that chat...", "I mentioned earlier...", "as we said...".

  • They ask about a project, person, decision, or specific term that might be in their history but is not in your current context.

  • They request context that seems "lost" between sessions or make an implicit reference to continuity ("I kept working on X", "the approach we discussed").

  • You need background on the user to answer well (what they do, their projects, their technical preferences).

BEFORE saying things like "I have no record", "I do not remember", "this is the first time I hear about that", "this is not in my memory" or similar, invoke this tool and check the results. The likelihood the data is indexed is high.

get_chatA

Fetch a specific conversation from history by uuid (with pagination).

USE when:

  • You have already identified a relevant chat (usually via search_chats) and need more context than the result snippet.

  • The user gave you a specific uuid and asked you to review it.

  • You are following a thread and need detail from a specific chat that appeared earlier in the conversation.

DO NOT use to discover new chats without searching first. To find chats by topic or keyword, use search_chats.

If the chat is long, call again with messages_offset to paginate (truncated response is indicated by truncated: true and total_messages).

list_recent_chatsA

Chronological browse of the user's most recent Claude.ai chats.

USE when:

  • The user asks about recent activity without a specific keyword ("what have I been up to", "which chats did I have this week", "catch up on what I have been thinking about").

  • You need general context on the topics the user has been working on before going deeper.

  • The user asks to explore without knowing exactly what to look for.

For targeted topic/keyword searches use search_chats. This tool is for chronological scanning, not for searching.

find_relatedA

Find chats related to a free-form context (not to a short query).

Different from search_chats: this tool accepts long input (a paragraph, file contents, what is being discussed right now) and returns semantically related chats. Internally uses pure vector search (no FTS) because for long inputs exact words weigh less than embedding similarity.

USE when:

  • You want "more like this": you have text in hand and want to see which history chats covered similar topics.

  • The user is pasting a paragraph / code / log and you need prior context automatically, without having to craft a keyword query.

  • You want to proactively suggest relevant chats without the user asking for them.

For short-keyword searches, use search_chats. For a recent chronological list, list_recent_chats.

index_terminal_sessionsA

Index the user's local Claude Code / terminal sessions into Memex now.

Runs the same incremental scan as memex ingest-claude-code: reads ~/.claude/projects/**/*.jsonl and ingests new or changed sessions (source claude_code) so they become searchable. Unchanged sessions are skipped, so this is cheap. The CURRENT session is captured up to what Claude Code has flushed to its transcript on disk.

USE when the user asks to index / capture / save their terminal or Claude Code work into their memory ("indexá esta charla", "index my sessions", "guardá esta conversación en memex"). To also send it to the user's other devices, follow with sync_now.

Returns: JSON with status and, on success, indexed (new sessions), unchanged, messages, chunks, errors.

sync_nowA

Sync Memex with the user's other devices now (paired peers + shared folder).

Runs an immediate two-way reconcile with every paired device and, if a shared sync folder is configured (the dual-boot mode), a file sync. It propagates what is ALREADY indexed locally, so to include the CURRENT terminal session, call index_terminal_sessions first.

USE when the user asks to sync / push / send their conversations to their other devices now ("sincronizá con mis dispositivos", "sync now", "mandá esto a la mac"). Requires sync to be enabled (memex setup --sync or memex sync enable).

Returns: JSON with status and, on success, peers (per-device pulled/pushed/ forks) and file_sync (folder pulled/forks), or a message explaining why nothing ran (disabled, or no target configured).

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/dioniipereyraa/memex'

If you have feedback or need assistance with the MCP directory API, please join our Discord server