Discover the Best AI Tools

Self-hosted log ingestion, semantic search, question answering and anomaly dashboards built around ClickHouse and Qdrant. Uses Groq or local Ollama, with optional Jev query classification; generated explanations require checking against the logs.

Claude Code Trace displays and tails local Claude Code JSONL sessions in desktop, web or terminal interfaces. Optional Jev analysis highlights repeated work and efficiency patterns after previewing and confirming a redacted payload. Its scores are model judgments, not measured monetary cost; viewing is local, but optional analysis sends trace content externally.

rlg combines structured Rust logging with log filtering, reporting and an MCP server that exposes log files to agents. It requires Rust 1.88+ to build; native system-log sinks are macOS/Linux-specific, with file or stdout output elsewhere.

Tokenectomy Razor filters framework noise from error traces and redacts recognizable credentials through a local Rust gateway and MCP server. It also offers caching and token limits; filtering and redaction must not be treated as complete leak prevention or a fix for provider quotas.

Local-first Rust TUI and CLI for auditing coding-agent sessions, including cost, tokens, latency, failures and health.

ClawMetry provides zero-config observability and governance for 30 AI agent runtimes, including Claude Code, OpenAI Codex, Hermes, OpenClaw, and 26 others. It displays live token costs, sessions, tool calls, and crons.

A tool for querying, summarizing, and tracing logs across multiple cloud platforms.

Logfire-mcp is an open-source tool for managing logs with AI capabilities.