Skip to main content
Glama
NacyZh

ScholarAgent RAG MCP Server

by NacyZh
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
PAPERS_DIRNoPath to the directory containing PDF papers.path/to/papers
ENABLE_HYDENoEnable HyDE query expansion.false
LLM_API_KEYNoAPI key for HyDE LLM.
HYDE_BACKENDNoBackend for HyDE: 'api' or 'vllm'.api
LLM_BASE_URLNoBase URL for HyDE LLM API.
LLM_MODEL_NAMENoModel name for HyDE.
BGE_OFFLINE_MODENoEnable offline mode for BGE models; no downloads.false
HYDE_API_TIMEOUTNoTimeout in seconds for HyDE API calls.60
BGE_AUTO_DOWNLOADNoAutomatically download missing BGE models from Hugging Face.true
BGE_M3_MODEL_PATHNoPath to the BGE-M3 embedding model../rag/models/bge-m3
RAG_SUMMARY_TOP_PNoTop-p for summarization.0.8
PAPER_PARSER_DEVICENoDevice for PDF parsing (e.g., cpu, cuda).cuda
RAG_SUMMARY_API_KEYNoAPI key for the summary API.
RAG_SUMMARY_BACKENDNoBackend for paper summaries: 'api' or 'vllm'.api
HYDE_VLLM_MODEL_PATHNoPath to local vLLM model for HyDE.
TOOL_MAX_QUERY_CHARSNoMaximum characters for tool queries.
BGE_M3_MODEL_REVISIONNoRevision of the BGE-M3 model.main
RAG_SUMMARY_API_MODELNoModel name used in API requests.Qwen/Qwen3-8B-AWQ
TOOL_MAX_RETURN_CHARSNoMaximum characters for tool return text.
RAG_SUMMARY_MODEL_NAMENoName of the summary model.Qwen/Qwen3-8B-AWQ
RAG_SUMMARY_MODEL_PATHNoPath to the local summary model directory../rag/models/Qwen3-8B-AWQ
RAG_SUMMARY_MODEL_REPONoHugging Face repository for the summary model.Qwen/Qwen3-8B-AWQ
BGE_RERANKER_MODEL_PATHNoPath to the BGE reranker model../rag/models/bge-reranker-v2-m3
RAG_SUMMARY_CONCURRENCYNoConcurrency for background summary jobs.1
RAG_SUMMARY_TEMPERATURENoTemperature for summarization.0.2
RAG_SUMMARY_API_BASE_URLNoBase URL for the OpenAI-compatible API for summaries.http://127.0.0.1:8001/v1
RAG_SUMMARY_OFFLINE_MODENoEnable offline mode for the summary model.false
RAG_SUMMARY_AUTO_DOWNLOADNoAutomatically download the summary model if missing.true
RAG_SUMMARY_ENFORCE_EAGERNoEnforce eager execution for vLLM.true
RAG_SUMMARY_CPU_OFFLOAD_GBNoCPU offload in GB for vLLM.0
RAG_SUMMARY_MODEL_REVISIONNoRevision of the summary model.main
RETRIEVAL_EVAL_RESULTS_DIRNoDirectory for retrieval evaluation results../rag/evaluation/results
BGE_RERANKER_MODEL_REVISIONNoRevision of the BGE reranker model.main
RAG_SUMMARY_API_MAX_RETRIESNoMaximum retries for API calls.2
RETRIEVAL_EVAL_DATASET_PATHNoPath to retrieval evaluation dataset../rag/evaluation/data/retrieval_eval.jsonl
RAG_SUMMARY_MAX_INPUT_TOKENSNoMaximum input tokens for summarization.4096
RAG_SUMMARY_MAX_OUTPUT_TOKENSNoMaximum output tokens for summarization.2048
RAG_SUMMARY_API_TIMEOUT_SECONDSNoTimeout in seconds for API calls.180
TOOL_JOB_STATUS_WAIT_MAX_SECONDSNoMaximum wait seconds for job status polling.
TOOL_JOB_STATUS_WAIT_STEP_SECONDSNoStep seconds for job status polling backoff.
RAG_SUMMARY_GPU_MEMORY_UTILIZATIONNoGPU memory utilization fraction for vLLM.0.75
TOOL_JOB_STATUS_WAIT_INITIAL_SECONDSNoInitial wait seconds for job status polling.

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

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityInactive
ResponsivenessNo issues

Resources