agent-trace and agenttrace
These are competitors offering overlapping functionality—both provide local-first tracing and replay of agent tool calls and prompts—though agent-trace focuses on capture/replay while AgentTrace emphasizes interactive step debugging through a web UI.
About agent-trace
Siddhant-K-code/agent-trace
strace for AI agents. Capture and replay every tool call, prompt, and response from Claude Code, Cursor, or any MCP client
It captures detailed session telemetry beyond LLM calls, including file operations, decision points, and error recovery, which can be exported to observability platforms like Datadog. The tool functions via injected hooks for full session capture in clients like Claude Code, an MCP stdio/HTTP proxy for general MCP clients, or direct Python decorators for custom agent implementations. It offers features like plain-English session summaries, token cost estimation, and diffing between sessions.
About agenttrace
Rxflex/agenttrace
AgentTrace is an open-source, local-first step debugger for AI agents. It provides a Python SDK for tracing your agent runs and a web UI to inspect spans, tool calls, prompts, and responses as an interactive tree.
Scores updated daily from GitHub, PyPI, and npm data. How scores work