Skip to main content
Glama
sportlotto-ux

unified-memory-mcp

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
HERMES_HOMENoBase for default paths (~/.hermes/*).~/.hermes
UM_DAG_FANINNoNodes of a level collapse into one above.5
UM_VEC_INDEXNoauto (build in reindex, KNN when dim matches) | off (always brute force).auto
UM_GRAPH_DECAYNomem_recall(hops>1): score multiplier per depth (depth-1).0.5
UM_LINK_FANOUTNomem_recall(hops>1): max links per node per direction.20
UM_ARCHIVE_PATHNoSeparate cold file.~/.hermes/unified_memory.archive.db
UM_ARCHIVE_BATCHNoHow many messages per archive pass.500
UM_DATABASE_PATHNoPath to the database.~/.hermes/unified_memory.db
UM_EMBEDDING_DIMNoSkip probe dim (openai), useful offline.
UM_CONTEXT_TOKENSNoEffective host context window.200000
UM_REDACT_ENABLEDNoSecrets gate on input (default ON — product is public).true
UM_RETENTION_DAYSNo0 = keep forever. >0 = hot data older than N days moves to archive once a week (lossless; deletion only manual purge).0
UM_SUMMARIZER_URLNoLLM summarization endpoint; without it extractive is used.
UM_ARCHIVE_SIZE_MBNoThreshold for hot DB: oldest messages move to archive.1024
UM_ASSEMBLY_BUDGETNoDefault token budget for mem_assemble.8000
UM_EMBEDDING_MODELNolocal: fastembed model strictly from registry; openai: passthrough name.paraphrase-multilingual-mpnet-base-v2
UM_RECALL_MAX_HOPSNomem_recall(hops): BFS traversal ceiling for links (above is error).3
UM_REDACT_PATTERNSNoSubset of catalog, comma-separated.api_key,bearer_token,password_assignment,private_key
UM_EVIDENCE_PARTIALNomem_evidence(cite): token coverage threshold for partial.0.5
UM_FRESH_TAIL_COUNTNoFresh messages never compressed.20
UM_SUMMARIZER_MODELNoModel for LLM summarization; without it extractive.
UM_COMPACT_THRESHOLDNoFraction of window that triggers compaction.0.35
UM_EMBEDDING_BACKENDNolocal (fastembed) | openai (8127/any OpenAI-compatible).local
UM_EMBEDDING_TIMEOUTNoHTTP timeout, seconds.30.0
UM_EVIDENCE_MAX_REFSNomem_evidence: max refs per call (excess go to rejections).50
UM_EMBEDDING_BASE_URLNoBase for backend=openai.http://127.0.0.1:8127
UM_EVIDENCE_MAX_CHARSNomem_evidence: how many chars of body to take from each ref.8000
UM_SUMMARIZER_API_KEYNoBearer token for the summarizer endpoint.

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