Skip to main content
Glama
ChrystianSchutz

ThreadShelf MCP Server

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
HOSTNoServer host. Set explicitly only when trusted LAN access is required.127.0.0.1
PORTNoServer port.3000
UPLOADS_DIRNoUploaded source files..uploads
LANCEDB_PATHNoLanceDB directory..lancedb
ALLOWED_HOSTSNoComma-separated extra Host/Origin names for trusted LAN access.
CHUNK_MAX_CHARSNoMax characters per embedded chunk.2000
COLLECTIONS_PATHNoManual-collections registry file..collections.json
EMBED_BATCH_SIZENoEmbedding batch size during ingest.25
LLAMA_CPP_SERVERNoAbsolute path to an existing llama-server executable.
LLAMA_CPP_THREADSNoCPU generation threads; -1 lets llama.cpp choose.-1
LLAMA_MODEL_PATHSNoExtra model roots (; on Windows, : on macOS/Linux).
LLAMA_CPP_BASE_URLNoExisting loopback-only llama.cpp server URL.
LLAMA_CPP_MAIN_GPUNoMain GPU index for applicable profiles.0
OPENROUTER_API_KEYNoOpenRouter key; may be set in .env, never exposed to the browser.
CHUNK_OVERLAP_CHARSNoOverlap between long chunks.100
MASTER_PROMPTS_PATHNoSaved master (system) prompts..threadshelf/master-prompts.json
OPENROUTER_BASE_URLNoOverride primarily intended for testing.https://openrouter.ai/api/v1
LLAMA_CPP_GPU_LAYERSNoExact layer offload for the hybrid profile.20
LLAMA_CPP_SPLIT_MODENoMulti-GPU split: layer or row.layer
GENERATION_CONFIG_PATHNoNon-secret Experimental Alpha generation settings..threadshelf/generation.json
LLAMA_CPP_ACCELERATIONNoauto, cpu, gpu, hybrid, or multi-gpu.auto
LLAMA_CPP_CONTEXT_SIZENoManaged local server context size.8192
LLAMA_CPP_TENSOR_SPLITNoOptional multi-GPU proportions, e.g. 3,1.
THREADSHELF_TOOLS_PATHNollama.cpp discovery/installer root..threadshelf/tools
LLAMA_CPP_FLASH_ATTENTIONNoFlash Attention: auto, on, or off.auto
THREADSHELF_DISABLE_DEFAULT_MODEL_PATHSNoSet 1 to scan only explicitly configured roots.0

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/ChrystianSchutz/ThreadShelf'

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