Skip to main content
Glama

LCBro

by lcbro
default.yaml•4.75 kB
server: host: "localhost" port: 3000 browser: engine: playwright # playwright | cdp headless: true defaultTimeoutMs: 30000 storageDir: /data/profiles maxContexts: 8 # Chrome DevTools Protocol (CDP) configuration cdp: enabled: false # enable CDP browser support host: "localhost" # CDP host port: 9222 # CDP port (default Chrome debug port) autoDetect: true # auto-detect running browsers maxRetries: 3 # connection retry attempts retryDelay: 1000 # delay between retries (ms) # Remote CDP server configuration remote: enabled: false # enable remote CDP server support url: null # remote CDP server URL (e.g., "https://cdp.example.com:9222") sslMode: "auto" # ssl mode: auto, enabled, disabled, insecure apiKey: null # API key for authentication (if required) headers: {} # additional headers for requests # Browser detection settings detection: enabled: true # enable browser detection ports: [9222, 9223, 9224, 9225, 9226] # ports to scan (for local detection) timeout: 5000 # detection timeout per port useRemote: false # use remote server for detection instead of local scanning # Browser launch settings (when auto-launching) launch: autoLaunch: false # auto-launch browser if none found browserPath: null # custom browser path (null = auto-detect) userDataDir: null # custom user data directory additionalArgs: [] # additional browser arguments # Connection settings connection: timeout: 30000 # connection timeout keepAlive: true # keep connection alive reconnect: true # auto-reconnect on disconnect maxReconnects: 5 # max reconnection attempts network: proxyDefault: null security: allowDomains: ["example.com", "gov.br"] blockPrivateNetworks: true llm: defaultModel: "ollama:llama3.1" # replace with your model: ollama:model or jan:model maxOutputTokens: 2000 temperature: 0 host: "localhost" # IP address of Ollama/JAN server port: 11434 # Ollama port (default 11434) janPort: 1337 # JAN port (default 1337) autoPreprocess: true # automatic data preprocessing (recommended) # Advanced preprocessing configuration preprocessing: enabled: true # enable/disable all preprocessing intelligentMode: true # use local LLM for prompt generation (recommended) fallbackToTemplates: true # fallback to templates if intelligent mode fails # Content size thresholds for auto-preprocessing thresholds: html: 3000 # auto-preprocess HTML above this size (bytes) text: 5000 # auto-preprocess text above this size (bytes) json: 1000 # auto-preprocess JSON above this size (bytes) # Model preferences for preprocessing (ordered by priority) preferredModels: - "ollama:qwen2.5:7b" # fastest for preprocessing - "ollama:llama3.2:3b" # very fast, small model - "ollama:mistral:7b" # good balance - "ollama:llama3.1:8b" # stable choice - "jan:llama-3.2-3b" # JAN fallback - "jan:mistral-7b" # JAN alternative # Analysis settings analysis: maxContentSample: 1000 # max chars to analyze for prompt generation maxAnalysisTokens: 300 # max tokens for content analysis analysisTemperature: 0.1 # low temperature for consistent analysis limits: maxChars: 300000 maxScreenshotBytes: 8000000 logging: level: info # Detailed LLM logging for debugging and statistics llm: enabled: true # enable detailed LLM logging logPrompts: true # log all prompts sent to LLM logResponses: true # log all responses from LLM logTokens: true # log token usage statistics logPerformance: true # log timing and performance metrics logPreprocessing: true # log preprocessing analysis and results # Data logging settings maxPromptLength: 2000 # max chars to log for prompts (truncate if longer) maxResponseLength: 1000 # max chars to log for responses (truncate if longer) maxInputDataLength: 5000 # max chars to log for input data (truncate if longer) # Performance tracking trackMetrics: true # track performance metrics metricsInterval: 100 # log metrics every N operations

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/lcbro/lcbro-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server