mcp-llm-bridge
by JNZader
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| HF_TOKEN | No | HuggingFace token for private repos in model discovery | |
| OLLAMA_URL | No | Ollama API endpoint | http://localhost:11434 |
| EMBEDDER_MODE | No | Embedding mode: 'local' uses Xenova transformer, 'api' forces OpenAI API | local |
| LM_STUDIO_URL | No | LM Studio API endpoint | http://localhost:1234 |
| OPENAI_API_KEY | No | Fallback API embedder key for semantic code search | |
| VOYAGE_API_KEY | No | Alternative API embedder key for semantic code search | |
| LATENCY_ROUTING | No | Enables latency-based routing (set to 'true') | |
| MCP_SERVERS_DIR | No | Directory for dynamic MCP server plugins | ./mcp-servers |
| LLM_GATEWAY_PORT | No | HTTP server port | 3456 |
| FALLBACK_STRATEGY | No | Enables free-model fallback routing (e.g., 'free-models') | |
| LOCAL_LLM_ENABLED | No | Enable local LLM routing (set to 'true') | false |
| FREE_MODEL_CATALOG | No | Loads the free-model catalog at startup (set to 'true') | |
| LLM_GATEWAY_DB_PATH | No | SQLite vault path | ~/.llm-gateway/vault.db |
| MCP_DYNAMIC_SERVERS | No | Enable dynamic MCP server plugins (set to 'true') | |
| AUTO_DISCOVER_MODELS | No | Enable HF auto-discovery of local models at startup | false |
| TRANSFORMERS_OFFLINE | No | Set to '1' to use only cached HuggingFace model, skip download | |
| MODEL_ROUTING_ENABLED | No | Enable model routing (set to 'true') | |
| LLM_GATEWAY_AUTH_TOKEN | No | Bearer token for HTTP routes, must be at least 32 characters | |
| LLM_GATEWAY_MASTER_KEY | No | 64-char hex key, otherwise auto-generated and saved to ~/.llm-gateway/master.key | |
| MAX_COMPARISON_COST_USD | No | Caps comparison-service spending | |
| ENABLE_OUTPUT_COMPRESSION | No | Enable RTK-style output compression | true |
| LLM_GATEWAY_AUTH_REQUIRED | No | Force auth on or off explicitly | |
| OPTIMIZE_MESSAGES_ENABLED | No | Enable three-part prompt optimization | true |
| LLM_GATEWAY_SECURITY_PROFILE | No | Security profile for MCP tool exposure: local-dev, restricted, or open | local-dev |
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 | |
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
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/JNZader/mcp-llm-bridge'
If you have feedback or need assistance with the MCP directory API, please join our Discord server