Faravex
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| FARAVEX_DIM | No | Embedding dimension: 768, 512, 256, or 128. | |
| FARAVEX_ROOT | No | Project root used by the server (default: current working directory). | cwd |
| FARAVEX_DTYPE | No | Model dtype; empty means bfloat16 on CUDA and float32 on CPU. float16 is never used. | |
| FARAVEX_DEVICE | No | Device for model inference: cuda or cpu. | |
| FARAVEX_BROWSER | No | Path to the Chrome or Edge executable used for browser recording. | |
| FARAVEX_MCP_PORT | No | MCP HTTP port when running in http mode (default: 8766). | 8766 |
| FARAVEX_HTTP_PORT | No | Browser UI HTTP port (default: 8765). | 8765 |
| FARAVEX_INDEX_DIR | No | Directory where indexes are stored (default: indexes/). | |
| FARAVEX_TEXT_ONLY | No | Set to 1 to skip vision/audio encoders and load only the text encoder. | |
| FARAVEX_BATCH_SIZE | No | Batch size for indexing/search (maximum 64). | |
| FARAVEX_DOC_PROMPT | No | Prompt prefix used when embedding documents/chunks. | |
| FARAVEX_EXTRA_EXTS | No | Additional file extensions to index, in addition to the built-in ones. | |
| FARAVEX_MODEL_PATH | No | Path to the EmbeddingGemma2 model folder (default: models/embeddinggemma). | |
| FARAVEX_CUDA_GRAPHS | No | Set to 1 to replay a captured CUDA graph for faster queries. | |
| FARAVEX_MAX_SEQ_LEN | No | Maximum model sequence length (default: 8192). | 8192 |
| FARAVEX_QUERY_PROMPT | No | Prompt prefix used when embedding queries. | |
| FARAVEX_TOKEN_BUDGET | No | Batch token budget; 0 means auto-select based on free VRAM. | 0 |
| FARAVEX_ALLOW_NETWORK | No | Set to 1 to disable the in-process block on non-loopback sockets and DNS. | |
| FARAVEX_CHUNK_MAX_CHARS | No | Maximum characters per chunk (default: 2400). | 2400 |
| FARAVEX_CHUNK_MAX_LINES | No | Maximum lines per chunk (default: 60). | 60 |
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 issues