MediaSearch
by bili14
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| WEMM_DATA | No | Direct path to the vector library file in stdio local mode; highest priority. | |
| WEMM_HOME | No | User directory root from which data/model/HF cache directories are derived when not otherwise set. | ~/.medisearch |
| WEMM_HOST | No | Host for the HTTP server (medisearch serve). | 127.0.0.1 |
| WEMM_PORT | No | Port for the HTTP server (medisearch serve). | 8686 |
| HF_ENDPOINT | No | Hugging Face mirror endpoint (e.g. https://hf-mirror.com) used by huggingface_hub for model downloads. | |
| WEMM_CONFIG | No | Path to config.json, overriding the default location. | <data_dir>/config.json |
| WEMM_BACKEND | No | Backend mode for stdio: auto (probe running server -> api, else local), api (force HTTP passthrough), or local (read local vector library and load local model). | auto |
| WEMM_HF_HOME | No | Hugging Face home directory used by the project. | ~/.medisearch/hf |
| WEMM_API_BASE | No | Base URL of the medisearch HTTP server used in api backend mode. | http://127.0.0.1:8686 |
| WEMM_DATA_DIR | No | Data directory containing embeddings.json, scopes.json, settings.json, config.json. Defaults to ./data/ in a repo checkout or ~/.medisearch/data for pip installs. | |
| WEMM_READ_ONLY | No | Set to 1 to enable read-only mode: rejects all write operations (adding/changing/removing scopes, building library, changing settings). Search and status remain available. | 0 |
| WEMM_LM_BED_API | No | LM Studio API endpoint probed before loading the model; if a model is already loaded there, loading is refused to avoid GPU contention. | http://127.0.0.1:1234/api/v0/models |
| WEMM_MODELS_DIR | No | Models directory root; superseded when model_path is set. | ~/.medisearch/models |
| WEMM_MODEL_PATH | No | Path to the local model weights directory. If weights are present they are used directly (offline capable); otherwise the model is auto-downloaded. | ~/.medisearch/models/WeMM-Embedding-2B |
| WEMM_MODEL_REPO | No | Hugging Face repository ID of the embedding model to download. | tencent/WeMM-Embedding-2B |
| WEMM_API_TIMEOUT | No | Timeout in seconds for search passthrough in stdio api mode (first search may be slow due to JIT warmup). | 300 |
| WEMM_FALLBACK_LIB | No | Fallback library file used when the main library is empty. | |
| WEMM_PROBE_TIMEOUT | No | Timeout in seconds for service probing in auto mode; may be increased if cold /status is slow. | 2.0 |
| WEMM_HF_MODULES_CACHE | No | Hugging Face modules cache directory. | ~/.medisearch/hf/modules |
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
| Name | Description |
|---|---|
No tools | |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
This server cannot be deployed
Maintenance
ActivityMaintained
ResponsivenessNo recent community issues