Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| host | No | Host to bind to (default: 127.0.0.1) | |
| port | No | Port to listen on (default: 39400) | |
| config | Yes | Path to the YAML configuration file | |
| log_level | No | Log level (DEBUG, INFO, WARNING, ERROR) | |
| no_meta_mcp | No | Disable Meta-MCP mode |
Capabilities
Features and capabilities supported by this server
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": true
} |
| resources | {
"subscribe": true,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| gateway_list_servers | List all 0 connected MCP backend servers with their status, tool count, and circuit-breaker state. |
| gateway_list_tools | List tools from a specific backend, or omit server to list all 0 tools across 0 backends. Returns names and descriptions — use gateway_search_tools for ranked results with full schemas. |
| gateway_search_tools | Search 0 tools across 0 servers by keyword. Returns ranked matches with full schemas, saving ~95% context tokens vs loading all tool definitions. Supports multi-word queries and synonym expansion. |
| gateway_invoke | Invoke any tool on any backend server. Routes through the gateway's auth, rate-limit, caching, and failsafe middleware. Use gateway_search_tools first to discover the right tool and server. |
| gateway_get_stats | Get usage statistics including invocations, cache hits, token savings, and top tools |
| gateway_cost_report | Return current session and API-key spend. Includes total cost, call count, and breakdown by backend and tool. Per-key totals are shown for 24 h / 7 d / 30 d rolling windows. |
| gateway_webhook_status | List registered webhook endpoints and their delivery statistics (received, delivered, failures, last event) |
| gateway_run_playbook | Execute a multi-step playbook (collapses multiple tool calls into one invocation) |
| gateway_kill_server | Immediately disable routing to a backend server (operator kill switch). The server's tools remain visible in search/list but are marked as disabled. |
| gateway_revive_server | Re-enable routing to a previously disabled backend server. Also resets the error budget so the server gets a clean slate. |
| gateway_set_profile | Switch the active routing profile for this session. A routing profile restricts which tools and backends are available. |
| gateway_get_profile | Show the active routing profile for this session and what it allows or denies. |
| gateway_list_disabled_capabilities | List capabilities that have been automatically disabled due to a high error rate. Each entry shows the backend, capability name, and how long it has been suspended. Disabled capabilities auto-recover after the configured cooldown period (default 5 min). Use gateway_revive_server to manually re-enable an entire backend immediately. |
| gateway_list_profiles | List all available routing profiles with their descriptions. Use gateway_set_profile to switch to a profile that narrows the visible toolset to the current task (e.g. "coding", "research"). |
| gateway_reload_config | Trigger an immediate reload of config.yaml from disk without restarting the gateway. Returns a summary of what changed (backends added/removed/modified, profile updates). Server host/port changes require a restart and are reported but not applied. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |