Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
HOSTNoServer bind address.0.0.0.0
PORTNoServer port.8000
BASE_URLNoOverride base URL for LLM.
BM25_RRF_KNoRRF constant K for hybrid search.60
CHAT_MODELNoChat completion model.openai/gpt-4o-mini
MINIO_HOSTNoMinIO endpoint (host:port).localhost:9000
OUTPUT_DIRNoTemporary directory for downloaded files.system temp
MAX_WORKERSNoWorkers for folder processing.3
BM25_ENABLEDNoEnable BM25 full-text search.true
MINIO_ACCESSNoMinIO access key.minioadmin
MINIO_BUCKETNoDefault bucket name.raganything
MINIO_SECRETNoMinIO secret key.minioadmin
MINIO_SECURENoUse HTTPS for MinIO.false
VISION_MODELNoVision model for image processing.openai/gpt-4o
EMBEDDING_DIMNoEmbedding vector dimension.1536
POSTGRES_HOSTNoPostgreSQL host.localhost
POSTGRES_PORTNoPostgreSQL port.5432
POSTGRES_USERNoPostgreSQL user.raganything
BRICKS_API_KEYNoAPI key for Bricks API authentication (publish).
MAX_TOKEN_SIZENoMax token size for embeddings.8192
ALLOWED_ORIGINSNoCORS allowed origins.["*"]
CLASSICAL_RRF_KNoRRF constant K for classical hybrid search.60
DOCUMENT_PARSERNoDocument parser for LightRAG pipeline.kreuzberg
EMBEDDING_MODELNoEmbedding model.text-embedding-3-small
BM25_TEXT_CONFIGNoPostgreSQL text search configuration.english
COSINE_THRESHOLDNoSimilarity threshold for vector search.0.2
RAG_STORAGE_TYPENoStorage backend: postgres or local.postgres
POSTGRES_DATABASENoPostgreSQL database name.raganything
POSTGRES_PASSWORDNoPostgreSQL password.raganything
UVICORN_LOG_LEVELNoUvicorn log level.critical
BRICKS_API_BASE_URLNoBricks platform base URL.https://analyse.bricks.co
BRICKS_BEARER_TOKENNoBearer token for Bricks API authentication (list documents).
OPEN_ROUTER_API_KEYYesRequired. OpenRouter API key for LLM access.
OPEN_ROUTER_API_URLNoOpenRouter base URL.https://openrouter.ai/api/v1
CLASSICAL_CHUNK_SIZENoMax characters per chunk for classical RAG.1000
MAX_CONCURRENT_FILESNoConcurrent file processing limit.1
BRICKS_PUBLISH_DRY_RUNNoWhen true, publish_section_version returns a payload preview.true
CLASSICAL_TABLE_PREFIXNoPrefix for PGVectorStore table names.classical_rag_
CLASSICAL_CHUNK_OVERLAPNoOverlap characters between chunks.200
ENABLE_IMAGE_PROCESSINGNoProcess images during indexing.true
ENABLE_TABLE_PROCESSINGNoProcess tables during indexing.true
CLASSICAL_LLM_TEMPERATURENoTemperature for LLM calls.0.0
ENABLE_EQUATION_PROCESSINGNoProcess equations during indexing.true
CLASSICAL_RELEVANCE_THRESHOLDNoMinimum LLM judge score.5.0
CLASSICAL_NUM_QUERY_VARIATIONSNoNumber of multi-query variations.3

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/SoluDevTech/mcp-raganything'

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