cedendahlkim/agentlens
🔍 X-ray vision for your AI agents - Real-time observability for MCP/LLM agents
Platform-specific configuration:
{
"mcpServers": {
"agentlens": {
"command": "npx",
"args": [
"-y",
"agentlens"
]
}
}
}Add the config above to .claude/settings.json under the mcpServers key.
<div align="center">
X-ray vision for your AI agents - Real-time observability for MCP/LLM agents
[](https://python.org) [](https://opensource.org/licenses/MIT) [](https://badge.fury.io/py/agentlens) [](https://github.com/gracestack-ab/agentlens)
Install • Quick Start • Demo • API • Contributing
</div>
---
AI agents are black boxes. When you run an MCP agent, you have no idea what's happening:
AgentLens gives you X-ray vision into your AI agents.
---
Wrap any MCP client with 2 lines of code and get a real-time dashboard:
from agentlens import wrap
# That's it! 🎉
client = wrap(your_mcp_client)
# All calls are now tracked in real-time
result = await client.call_tool("search", {"query": "AI agents"})*Live dashboard showing real-time tool calls, latency, tokens, and costs*
---
Loading reviews...