Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active
TypeScript aianthropicartifactsawsazurechatgptchatgpt-cloneclaudeclonedeepseekgeminigooglegpt-5librechatmcpo1openairesponses-apivisionwebui
import path from 'path';
import type winston from 'winston';
/**
* Determine the log directory in a cross-compatible way.
* Priority:
* 1. LIBRECHAT_LOG_DIR environment variable
* 2. If running within LibreChat monorepo (when cwd ends with /api), use api/logs
* 3. If api/logs exists relative to cwd, use that (for running from project root)
* 4. Otherwise, use logs directory relative to process.cwd()
*
* This avoids using __dirname which is not available in ESM modules
*/
export const getLogDirectory = (): string => {
if (process.env.LIBRECHAT_LOG_DIR) {
return process.env.LIBRECHAT_LOG_DIR;
}
const cwd = process.cwd();
// Check if we're running from within the api directory
if (cwd.endsWith('/api') || cwd.endsWith('\\api')) {
return path.join(cwd, 'logs');
}
// Check if api/logs exists relative to current directory (running from project root)
// We'll just use the path and let the file system create it if needed
const apiLogsPath = path.join(cwd, 'api', 'logs');
// For LibreChat project structure, use api/logs
// For external consumers, they should set LIBRECHAT_LOG_DIR
if (cwd.includes('LibreChat')) {
return apiLogsPath;
}
// Default to logs directory relative to current working directory
return path.join(cwd, 'logs');
};
/**
* The level set every logger in this package is built with. Winston resolves a
* level name against this map, so a name outside it resolves to `undefined` and
* the transport drops every message.
*/
export const logLevels: winston.config.AbstractConfigSetLevels = {
error: 0,
warn: 1,
info: 2,
http: 3,
verbose: 4,
debug: 5,
activity: 6,
silly: 7,
};
/**
* Console verbosity for every winston logger in this package.
*
* Each logger's Console transport used to hard-code `level: 'info'`, which
* silently overrides the logger's own level — winston resolves a transport's
* explicit level ahead of its parent's. `CONSOLE_LOG_LEVEL` makes that
* verbosity configurable, and `silent` turns console output off entirely.
*
* An unrecognized name falls back rather than reaching winston: it would
* resolve to `undefined` there and silently discard every console log, so a
* typo would read as a dead deployment. Say so on the way past — the logger
* cannot report its own misconfiguration.
*/
export const resolveConsoleLevel = (fallback = 'info'): { level: string; silent: boolean } => {
const requested = process.env.CONSOLE_LOG_LEVEL?.trim().toLowerCase();
if (!requested) {
return { level: fallback, silent: false };
}
if (requested === 'silent') {
return { level: fallback, silent: true };
}
if (Object.prototype.hasOwnProperty.call(logLevels, requested)) {
return { level: requested, silent: false };
}
console.warn(
`[logger] Ignoring unknown CONSOLE_LOG_LEVEL "${process.env.CONSOLE_LOG_LEVEL}"; using "${fallback}". ` +
`Expected one of: ${Object.keys(logLevels).join(', ')}, silent.`,
);
return { level: fallback, silent: false };
};