Skip to main content
Glama
james-choncholas

matrix-nio-mcp

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
MCP_PORTNoPort for the HTTP server8000
QDRANT_HOSTNoQdrant hostname (qdrant inside Docker Compose)localhost
QDRANT_PORTNoQdrant port6333
WEBHOOK_URLNoOpenAI-compatible base URL for the LLM callback
IGNORED_ROOMSNoComma-separated list of Matrix room IDs to exclude
WEBHOOK_MODELNoModel name passed to the LLMgpt-4o-mini
WEBHOOK_TOOLSNoOptional JSON string of tools/parameters merged into the chat completions request body
BACKFILL_LIMITNoMessages fetched per page per room during startup backfill100
MATRIX_USER_IDYesFull MXID, e.g. @bot:example.org
OPENAI_API_KEYYesOpenAI API key for embeddings
EMBEDDING_MODELNoOpenAI embedding modeltext-embedding-3-small
HTTP_AUTH_TOKENNoIf set, requires Authorization: Bearer <token> for all endpoints
MATRIX_DEVICE_IDYesDevice ID — must be stable across restarts for E2EE
MATRIX_STORE_PATHNoPath for the Olm E2EE crypto database (created if absent)~/.cache/nio-mcp/store
QDRANT_COLLECTIONNoQdrant collection namematrix_messages
SSE_QUEUE_MAXSIZENoPer-subscriber SSE event queue cap (oldest dropped when full)100
ALLOW_SEND_MESSAGENoSet to true to enable the send_message toolfalse
BACKFILL_PAGES_MAXNoMaximum backfill pages per room; 0 = full history10
MATRIX_ACCESS_TOKENYesLong-lived access token
MESSAGE_BUFFER_SIZENoIn-memory ring buffer size for get_recent_messages500
WEBHOOK_BEARER_TOKENNoBearer token sent in the Authorization header
EMBEDDING_VECTOR_SIZENoOutput dimension requested from the model and used for the Qdrant collection1536
MATRIX_HOMESERVER_URLYesHomeserver URL, e.g. https://matrix.example.org
WEBHOOK_PROMPT_HEADERNoText prepended once before all per-message linesNew Matrix messages:
MATRIX_KEY_BACKUP_FILENoPath to an Element-exported E2EE key file; see Decrypting historical messages in encrypted rooms
MATRIX_SYNC_TIMEOUT_MSNoMatrix /sync long-poll timeout in milliseconds30000
WEBHOOK_PROMPT_PER_MSGNoTemplate rendered once per buffered message{sender_name} ({sender}) in {room_name} ({room}): {message}
WEBHOOK_COOLDOWN_SECONDSNoSeconds of silence before the LLM is called300
MATRIX_KEY_BACKUP_PASSPHRASENoPassphrase chosen when exporting; required when MATRIX_KEY_BACKUP_FILE is set

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/james-choncholas/matrix-nio-mcp'

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