Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import pino, { type DestinationStream } from 'pino'
export const REDACT_PATHS = [
'accessToken',
'refreshToken',
'chatgptAccountId',
'accountId',
'executionTicket',
'authorization',
'headers.authorization',
'req.headers.authorization',
'attemptReceipt',
'id_token',
'idToken',
'pkceVerifier',
'refresh_token',
]
export function createProxyLogger(destination?: DestinationStream) {
return pino(
{
level: destination
? 'info'
: (process.env.LOG_LEVEL ?? (process.env.NODE_ENV === 'test' ? 'silent' : 'info')),
base: { svc: 'codex-llm-proxy' },
timestamp: pino.stdTimeFunctions.isoTime,
redact: { paths: REDACT_PATHS, remove: true },
},
destination
)
}
export const logger = createProxyLogger()