Loggles
Provides a local-first log sink for development agents to query structured log data efficiently, transforming coding agents into runtime companions.
Loggles is designed as a local-first log sink that seamlessly integrates with your coding agents, empowering them to act as intelligent runtime companions. It eliminates the need for manual copy-pasting of log output into AI tools by allowing agents like Claude Code or Cursor to directly query structured log data. Agents can filter logs precisely by service, level, trace ID, or time window, receiving structured insights for advanced reasoning without token waste, making bug investigation and runtime behavior observation more intuitive and effective.
주요 기능
01Ingest logs via OTLP HTTP/protobuf from various languages (e.g., .NET, Node.js, Python, Go, Java, Ruby, PHP).
02Comprehensive MCP (Multi-Modal Companion Protocol) tools for AI agents to search, filter, and analyze logs.
03Supports customizable log retention, authentication via API keys, and local OAuth 2.0 + PKCE for MCP clients.
04Flexible storage options, including SQLite for simplicity and PostgreSQL for scalability.
05Integrated Claude Code skill for investigative debugging and exploratory runtime observation.
067 GitHub stars
사용 사례
01Investigating bugs: Allow AI agents to trace failures through logs, cross-reference with source code, and identify root causes.
02Watching runtime behavior: Enable AI agents to tail live logs, trace requests by ID, and narrate application actions without attaching a debugger.
03Auditing log quality: Generate reports on instrumentation coverage, identify missing templates, and detect unstructured messages.