SpanHound MCP
SpanHound captures local AI agent traces via OpenTelemetry and exposes read-only MCP tools to inspect failures, approval decisions, token usage, and run comparisons. It supports Python agents with optional auto-instrumentation or existing OTLP exporters, allowing users to query JSONL spans and OTLP JSON exports for debugging and regression analysis.
Package and access
Local package with source-attributed installation metadata. Koskamo does not execute packages during review.
What you can do with it
- Inspect agent run failures and approval decisions
- Analyze token usage across agent runs
- Run regressions on recorded agent traces
- Query OpenTelemetry traces from local agent calls
Source and maintenance
- Maintenance checked
Packages and installation
Source-provided package data. Koskamo has not executed or live-probed this package.
abhishekash-mcp-trace 0.1.1
{
"version": "0.1.1",
"transport": "stdio",
"identifier": "abhishekash-mcp-trace",
"runtimeHint": "uvx",
"registryType": "pypi",
"registryBaseUrl": "https://pypi.org"
}Tools
Reported by the attributed source; not independently observed by Koskamo.
Connect from your client
Each snippet configures the same source-provided package; replace placeholders before use.
Cursor
{
"mcpServers": {
"spanhound-mcp": {
"command": "uvx",
"args": [
"abhishekash-mcp-trace==0.1.1"
]
}
}
}Cline
{
"mcpServers": {
"spanhound-mcp": {
"command": "uvx",
"args": [
"abhishekash-mcp-trace==0.1.1"
]
}
}
}VS Code
{
"servers": {
"spanhound-mcp": {
"type": "stdio",
"command": "uvx",
"args": [
"abhishekash-mcp-trace==0.1.1"
]
}
}
}Claude Desktop
{
"mcpServers": {
"spanhound-mcp": {
"command": "uvx",
"args": [
"abhishekash-mcp-trace==0.1.1"
]
}
}
}Let your agent find its tools
Find servers like SpanHound MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.