Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
PORTNoHTTP server port.3100
DOCKERNoDocker defaults: MEMORIA_DIR becomes /data/memoria and auto-generating an on-disk encryption key logs a warning.false
BIND_ALLNoSet to true to bind to 0.0.0.0 instead of 127.0.0.1, required for Docker and Cloud Run.false
MEMORIA_DIRNoRoot Memoria directory. Defaults to /data/memoria when DOCKER=true, otherwise ~/.memoria.
OPENAI_API_KEYNoAPI key for OpenAI embeddings. On its own it does not enable remote embeddings; MEMORIA_EMBEDDINGS=openai is also required.
MEMORIA_API_KEYNoBearer token for HTTP transport authentication for /mcp, /ingest, and /dashboard/api. Required for HTTP transport, distinct from the OAuth client secret.
MEMORIA_REPO_URLNoRepository URL shown in the dashboard's About card.https://github.com/Agripp87/memoria_mcp
MEMORIA_EMBEDDINGSNoEmbedding provider: auto (local MiniLM if installed, else hash; never remote), openai (explicit opt-in, sends content to OpenAI), minilm (local all-MiniLM-L6-v2), or hash (n-gram lexical fallback).auto
MEMORIA_PUBLIC_URLNoPins the OAuth issuer and endpoint metadata base URL. Recommended in any network-exposed deployment.
MEMORIA_MODEL_CACHENoDirectory for the cached MiniLM model.
MEMORIA_RAW_ARCHIVENoEnable the durable, immutable provenance archive of ingested events under data/raw/.true
MEMORIA_TRUST_PROXYNoNumber of proxy hops in front of the HTTP server, used to find the client IP for rate limiting.1
MEMORIA_AUTO_COMPILENoSet to true to autonomously compile daily-log data into linked entity pages on a timer.false
MEMORIA_TOKEN_DB_DIRNoDirectory for the OAuth token database tokens.sqlite. Defaults to the data directory.
MEMORIA_AUTO_OPTIMIZENoSet to true to run index decay, promotion, and staleness passes on a timer.false
MEMORIA_LOCK_TAKEOVERNoWhen true, a server takes over a store lock held by a live server on another host instead of waiting for it to stop.false
MEMORIA_MODEL_OFFLINENoSet to true to forbid model downloads and use only the cached or pre-baked model.false
MEMORIA_ENCRYPTION_KEYNoAES-256-GCM master key as 64 hex characters for the collector's encrypted buffer and credential store. Required when BIND_ALL=true.
MEMORIA_OAUTH_CLIENT_IDNoOAuth client id the token endpoint expects and /register hands out.memoria
MEMORIA_VECTOR_SCAN_CAPNoMaximum chunks scanned per query for semantic candidate selection.5000
MEMORIA_RAW_ARCHIVE_BATCHNoNumber of buffered raw-archive records to flush after.20
MEMORIA_FILE_WATCHER_ROOTSNoComma-separated directories that file_watcher collector sources may watch. Unset or empty refuses every watcher.
MEMORIA_ALLOW_SHELL_SOURCESNoSet to true on trusted local instances only to permit shell_command custom sources.false
MEMORIA_COMPILE_INTERVAL_MSNoInterval for the automatic entity compile pass, in milliseconds; only used when MEMORIA_AUTO_COMPILE=true.86400000
MEMORIA_INGEST_MAX_ATTEMPTSNoFailed ingest attempts before a poison event is dead-lettered.3
MEMORIA_OAUTH_CLIENT_SECRETNoOAuth client secret. Required when BIND_ALL=true. Keep it distinct from the API key.
MEMORIA_REINDEX_INTERVAL_MSNoInterval for the periodic reindex sweep, in milliseconds.300000
MEMORIA_OPTIMIZE_INTERVAL_MSNoInterval for the automatic optimize pass, in milliseconds; only used when MEMORIA_AUTO_OPTIMIZE=true.86400000
MEMORIA_RAW_ARCHIVE_FLUSH_MSNoMilliseconds after which buffered raw-archive records are flushed.60000
MEMORIA_ALLOWED_REDIRECT_HOSTSNoComma-separated hostnames permitted as OAuth redirect_uri targets. HTTPS is required except for localhost and 127.0.0.1.claude.ai,claude.com,anthropic.com,localhost,127.0.0.1
MEMORIA_BACKPRESSURE_THRESHOLDNoFraction of buffer capacity at which source polling pauses so ingestion can drain the backlog.0.8
MEMORIA_REQUIRE_ENCRYPTION_KEYNoRefuse to start without MEMORIA_ENCRYPTION_KEY; no on-disk key fallback.false
MEMORIA_INSECURE_ALLOW_FALLBACKSNoEscape hatch allowing BIND_ALL=true to start without MEMORIA_OAUTH_CLIENT_SECRET, MEMORIA_ENCRYPTION_KEY, and MEMORIA_PUBLIC_URL.false
MEMORIA_RAW_ARCHIVE_ROTATE_BYTESNoByte size at which the raw archive rotates to a new part.5242880
MEMORIA_AUTHORIZE_REQUIRE_APPROVALNoWhen true, /authorize additionally requires operator approval via the API key or a dashboard login and consent page, and disables the client_credentials grant.false

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

Server capabilities have not been inspected yet.

Tools

Functions exposed to the LLM to take actions

NameDescription

No tools

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources