mcpbeat Sign in

Tracehub MCP Server

local only

Tracehub MCP runs on your own machine — the client starts it, so there is no endpoint to ping. 653 installs a week from pypi. Last commit 31 Oct 2026.

Query OTel traces across Jaeger, Tempo, Traceloop, Datadog, Sentry, X-Ray, New Relic, Honeycomb.

Installs per day peak 396 · avg 91 · +118% w/w
a month agotoday
653
Installs / week
pypi · tracehub-mcp
0
Stars
8 open issues
31 Oct 2026
Last commit
from GitHub
Apache-2.0
License
Python

Nothing serious here today

Today is the operative word: we check Tracehub MCP every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.

Three servers free · no card

Connect this server

This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.

run in your terminal
claude mcp add tracehub-mcp -e 'BACKEND_TYPE=<BACKEND_TYPE>' -e 'BACKEND_URL=<BACKEND_URL>' -- uvx tracehub-mcp
~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "tracehub-mcp": {
      "command": "uvx",
      "args": [
        "tracehub-mcp"
      ],
      "env": {
        "BACKEND_TYPE": "<BACKEND_TYPE>",
        "BACKEND_URL": "<BACKEND_URL>"
      }
    }
  }
}
~/.codex/config.toml
[mcp_servers.tracehub-mcp]
command = "uvx"
args = ["tracehub-mcp"]
env = { BACKEND_TYPE = "<BACKEND_TYPE>", BACKEND_URL = "<BACKEND_URL>" }
.cursor/mcp.json
{
  "mcpServers": {
    "tracehub-mcp": {
      "command": "uvx",
      "args": [
        "tracehub-mcp"
      ],
      "env": {
        "BACKEND_TYPE": "<BACKEND_TYPE>",
        "BACKEND_URL": "<BACKEND_URL>"
      }
    }
  }
}
.vscode/mcp.json
{
  "servers": {
    "tracehub-mcp": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "tracehub-mcp"
      ],
      "env": {
        "BACKEND_TYPE": "<BACKEND_TYPE>",
        "BACKEND_URL": "<BACKEND_URL>"
      }
    }
  }
}

Put your own values in place of <BACKEND_TYPE> (Which observability backend to query), <BACKEND_URL> (Backend API endpoint, e.g. http://localhost:16686 for Jaeger, https://api.datadoghq.com for Datadog, https://api.newrelic.com/graphql for New Relic, or https://api.honeycomb.io for Honeycomb. For the X-Ray backend this is decorative only (e.g. https://xray.us-east-1.amazonaws.com) - never dereferenced for a live request). The author’s registry entry marks them as required.

Optional settings from the registry entry, left out of the block above: BACKEND_API_KEY (API key or auth token. Required for Traceloop, Datadog, Sentry, New Relic, and Honeycomb; not used by Jaeger, Tempo, or X-Ray. For New Relic this must be a User API key, not an Ingest or License key; for Honeycomb this must be a Configuration Key, not an Ingest or Management key), BACKEND_AWS_REGION (AWS region, e.g. us-east-1. Required for the X-Ray backend only - X-Ray authenticates via boto3's standard AWS credential chain, not BACKEND_API_KEY), BACKEND_NEWRELIC_ACCOUNT_ID (New Relic account ID, required for the New Relic backend only - NerdGraph queries are user-scoped, not account-scoped, so this identifies which account's data to read), BACKEND_HONEYCOMB_DATASET (Honeycomb dataset slug, required for the Honeycomb backend only - the Query Data API is dataset-scoped), BACKEND_APP_KEY (Datadog Application key, required in addition to BACKEND_API_KEY for the Datadog backend only), BACKEND_TEMPO_INSTANCE_ID (Grafana Cloud stack/instance ID, used with BACKEND_API_KEY for Basic Auth (Tempo backend only, required for Grafana Cloud-hosted Tempo, not needed for self-hosted Tempo)), BACKEND_SENTRY_ORG (Sentry organization slug, required for the Sentry backend only), BACKEND_SENTRY_PROJECT (Sentry project slug, optional for the Sentry backend - narrows queries to one project), BACKEND_ENVIRONMENTS (Comma-separated environments for the Traceloop backend only), BACKEND_TIMEOUT (Request timeout in seconds), BACKEND_MAX_RESPONSE_BYTES (Ceiling on the decoded size of any single backend response, in bytes (1 to 1073741824). An oversized body is refused before it is buffered or parsed), LOG_LEVEL (Logging level), MAX_TRACES_PER_QUERY (Maximum traces validated per query (1-1000)), SLOW_REQUEST_THRESHOLD_MS (Logs a WARNING for any backend request slower than this, independent of LOG_LEVEL. Unset disables slow-request logging), QUERY_CACHE_TTL_SECONDS (Cache backend query results for this many seconds with in-flight request coalescing; finite, > 0 and at most 86400 (one day). Unset disables the cache), QUERY_CACHE_MAX_ENTRIES (Upper bound on the number of cached query results held at once (1-100000); the least-recently-used entry is evicted first), OTEL_EXPORTER_OTLP_ENDPOINT (Enables opt-in OTel self-instrumentation of tool calls when set. Unset means zero overhead - no TracerProvider is configured and no middleware is registered), OTEL_SERVICE_NAME (Service name reported in self-instrumentation spans, only used when OTEL_EXPORTER_OTLP_ENDPOINT is set), MCP_INCLUDE_ARGS_IN_SPANS (Include tool call arguments/results as OTel span attributes when self-instrumentation is enabled. Off by default since they may contain sensitive data).

Alternatives to Tracehub MCP

same job, measured the same way
New Relic MCP Server
by newrelic

Access New Relic observability data through MCP - query metrics, logs, traces, entities, and more

answering
Honeycomb
by honeycomb

Query Honeycomb observability data: traces, events, metrics, SLOs, triggers, and boards.

answering
OtlpDashboard
by adiakys

Query OpenTelemetry logs, traces and metrics from a self-hosted OTLP dashboard

local only
D
DataRunner
by datarunner

Query, monitor and schedule SQL across your databases: schemas, read-only SQL, alerts, hosts.

73 installs/wk local only
Spoor
by asinha07

What changed across metrics, logs and traces, and where it started (Grafana, Victoria, New Relic).

local only
MCP Trace
by abhishekash

Query OpenTelemetry traces from agent runs through MCP.

196 installs/wk local only
UserTapes
by mcdca

Query real user session replay data: tapes, transcripts, error/rage-click filters, alerts.

answering
Shopify Catalog Monitor
by dizzyvinci

Track price drops, stock-outs, restocks, and new/removed products across Shopify stores.

answering

Tracehub MCP — questions

Answers built from our own checks of this server.

Why is there no uptime for Tracehub MCP?
Tracehub MCP runs on your own machine over stdio — there is no network address to reach, so uptime cannot be measured for it by anyone. What can be measured is adoption: the pypi package tracehub-mcp was installed 653 times last week.
How do I connect Tracehub MCP?
Copy the ready config from this page. We generate it for Claude Code, Claude Desktop, Codex, Cursor and VS Code, each in the form that client actually reads. It runs locally, so the command pulls tracehub-mcp straight from pypi; nothing to host, nothing to sign up for.
How many people use Tracehub MCP?
The pypi package tracehub-mcp was installed 653 times in the last week. Week over week that is +118%. We show installs rather than GitHub stars on purpose: a star is a bookmark, an install is someone actually running it.
Is Tracehub MCP open source?
Yes — it is published under the Apache-2.0 licence, written in Python, 0 stars on GitHub and 8 open issues. The source link is on this page, so you can read exactly what it does with your data before you connect it.