Skip to main content
Glama
inkiebeard

vkb

by inkiebeard
OverviewSchemaRelated ServersScoreDiscussions
TypeScript
Hybrid

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
TLS_KEYNoPath to TLS private key file
MCP_PORTNoMCP server port (0 = stdio mode only, no auth layer)3333
OBS_PORTNoObservability HTTP/WebSocket server port4242
TLS_CERTNoPath to TLS certificate file (enables HTTPS/WSS on all servers)
EMBED_DIMNoEmbedding dimension (must match model output)768
LLM_MODELNoLLM model for relation extraction and summarisationgemma4:e4b
LOG_LEVELNoLog verbosity: debug, info, warn, errorinfo
CHUNK_SIZENoTarget chunk size in tokens512
MCP_SECRETNoBearer token for the HTTP MCP endpoint on MCP_PORT (leave unset to disable)
OBS_SECRETNoBearer token for the REST/browser API on OBS_PORT (leave unset to disable)
EMBED_MODELNoEmbedding modelnomic-embed-text
DATABASE_URLNoPostgres connection stringpostgres://localhost/vkb
JOB_TTL_DAYSNoDays before completed/failed jobs are expired7
CHUNK_OVERLAPNoOverlap between adjacent chunks64
IVFFLAT_LISTSNoNumber of ivfflat lists100
RAWSTORE_PATHNoRoot path for filesystem rawstore./rawstore
RELATION_TOP_KNoNearest neighbours considered per chunk10
OLLAMA_BASE_URLNoOllama API base URLhttp://localhost:11434
RAWSTORE_ADAPTERNoRaw content storage: filesystem or s3filesystem
RETUNE_SUMMARISENoRegenerate entity summaries during retunefalse
SECTION_MAX_SIZENoMax chunks per section8
SECTION_STRATEGYNoSectioning strategy: similarity_valley or positionalsimilarity_valley
IVFFLAT_THRESHOLDNoChunk count above which an ivfflat index is created/maintained1000
RELATION_TTL_DAYSNoDays before unconfirmed relations are pruned30
INGEST_MAX_RETRIESNoMax times a failed job is re-queued3
RAWSTORE_S3_BUCKETNoS3 bucket name (when RAWSTORE_ADAPTER=s3)
RELATION_THRESHOLDNoMinimum cosine similarity to create a relation0.75
WORKER_CONCURRENCYNoNumber of ingest worker processes2
SECTION_WINDOW_SIZENoSliding window size for valley detection5
SUMMARY_CONCURRENCYNoParallel LLM calls during the summarising stage4
SUMMARY_PROMPT_FILENoEntity summary prompt file path
RAWSTORE_S3_ENDPOINTNoS3-compatible endpoint URL (optional override)
RETUNE_INTERVAL_HOURSNoHow often the retune worker runs automatically (0 = disabled)6
LLM_EXTRACT_CANDIDATESNoCandidate chunks the LLM considers per extraction pass20
LLM_RELATION_EXTRACTIONNoUse LLM to extract relations (set false to use heuristics only)true
SECTION_SPLIT_THRESHOLDNoCosine similarity drop that triggers a section boundary0.65
SUMMARY_MAX_INPUT_CHARSNoMax characters fed to the entity-level summary prompt (~3k tokens)12000
RELATION_CONFIDENCE_STEPNoIncrement applied on relation confirmation0.05
RELATION_PRUNE_THRESHOLDNoConfidence below which relations are pruned on retune0.6
CHUNK_SUMMARY_PROMPT_FILENoPer-chunk summary prompt file path
SECTION_SUMMARY_PROMPT_FILENoSection summary prompt file path
RELATION_EXTRACT_PROMPT_FILENoRelation extraction prompt file path

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

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/inkiebeard/galactic-vkb'

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