LogCore

A production-ready structured logging library for Python.

LogCore wraps Python’s stdlib logging with a single ergonomic entrypoint, structured JSON output, async-safe correlation IDs, automatic sensitive-data redaction, OpenTelemetry integration, and intelligent log sampling — all with zero required dependencies.

from logcore import get_logger

log = get_logger("myapp", json=True)
log.info("user login", user="alice", role="admin")

Getting started

• Quickstart
  • Install
  • A logger in two lines
  • Configuration
  • Five things you’ll actually use
  • What’s next

Guides

• Correlation IDs
  • Three ways to set a correlation ID
  • Reading the current correlation ID
  • Async safety
  • Output
  • See also
• Log sampling
  • When to use each strategy
  • Quickstart
  • How tail-based works
  • Using set_correlation_id instead of the context manager
  • Environment variables
  • Observability: stats()
  • What’s deliberately out of scope (for now)
  • See also
• OpenTelemetry integration
  • Install
  • Usage
  • Compatible backends
  • With correlation IDs
• Framework integration
  • Flask
  • FastAPI
  • ASGI middleware (generic)
• Capturing third-party logs
  • configure_stdlib
  • Using dictConfig instead
  • Correlation IDs across a request
  • Non-blocking delivery
• Sensitive data redaction
  • Default protected fields
  • What “partial masking” means
  • Customizing
  • What gets walked

Reference

• API reference
  • Top-level exports
  • Logger
  • Sampling
  • Correlation IDs
  • Timing helpers
  • Configuration
  • Formatters
  • Handlers
  • Lifecycle
  • Standard library interop
  • Web middleware
• Configuration reference
  • Core options
  • Output and delivery options
  • Sampling options
  • Boolean parsing for env vars
  • Invalid environment values
  • Logger caching and reconfiguration
  • Reading the current correlation ID without a logger
• Changelog
  • Unreleased
  • 0.1.7 - 2026-08-15
  • 0.1.6 - 2026-05-27
  • 0.1.5 - 2026-05-20
  • 0.1.4 - 2025-01-15
  • 0.1.3 - 2024-12-01
  • 0.1.2 - 2024-11-15
  • 0.1.1 - 2024-10-20
  • 0.1.0 - 2024-10-01

At a glance

• Single entrypoint. get_logger(name) returns a configured, cached logger — no dictConfig, no handler/formatter wiring.

• Structured by default. JSON output mode produces records that log aggregators can query directly.

• Async-safe correlation IDs. Per-task isolation via contextvars.

• Automatic redaction. Sensitive fields (passwords, tokens, keys) are partially masked — enough to confirm a value was present without leaking it.

• OpenTelemetry-aware. trace_id and span_id injected automatically when a span is active.

• Tail-based sampling. Buffer records per request, flush them on error, drop them on success. Pay nothing for healthy traffic; keep everything when something breaks.

• Strict typing. PEP 561–compliant py.typed marker; mypy-clean.

Install

pip install logcore

Optional extras:

pip install "logcore[colors]"  # colored console output
pip install "logcore[otel]"    # OpenTelemetry integration

Where to go next

• New to LogCore? Start with the quickstart.

• High-volume service? See log sampling.

• Need request-scoped logging? See correlation IDs.

• Looking for a specific function? See the API reference.