Skip to content
Memfault MCP Server

Memfault MCP Server

by Memfault (Nordic Semiconductor)
Added to Onei
Docs
Cloud & Infrastructure

Summary

Remote MCP server for Memfault (Nordic Semiconductor) exposing 18 read-only tools over device fleet telemetry — device vitals, timeseries, crash issues, traces, metrics and software versions.

Features

  • 18 read-only tools over device fleet telemetry
  • Device vitals, attributes, timeseries and reboot history
  • Crash issue search plus trace retrieval
  • Metrics and per-software-version vitals
  • OAuth 2.0 with PKCE, scoped to one org and project
  • Per-plan rate limits from 30 to 300 calls per hour

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

  1. Open your agent's MCP configuration file.
  2. Merge the snippet below into it.
  3. Restart the agent — the "Memfault MCP Server" tools will be available.
{
  "mcpServers": {
    "memfault-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://app.memfault.com/mcp"
      ]
    }
  }
}

Description

Memfault MCP Server is the hosted Model Context Protocol endpoint for Memfault, the embedded device observability platform now part of Nordic Semiconductor. It puts a connected-device fleet's telemetry behind an AI client, so questions like "which firmware version is rebooting most this week" can be answered without hand-building a dashboard.

Read-only by design. All 18 tools are reads, and the default OAuth scopes (device:read, issue:read, chart:read, deployment:read, alert:read, project:read, org:read, me:read) grant nothing else. An agent can investigate a fleet but cannot deploy to it.

Tool groups

  • Devices — device_get, device_getAttributes, device_getTimeseries, device_getVitals, device_search, device_listReboots
  • Issues and traces — issues_list, issue_search, issue_get, issue_getTraces, trace_get
  • Metrics — metrics_list, metrics_listReportTypes
  • Software versions — softwareVersions_list, softwareVersion_getVitals
  • Processing logs — processingLog_listEventTypes, processingLog_search
  • Projects — projects_list

Access and limits

Authentication is OAuth 2.0 with PKCE; a token is scoped to one organisation and project chosen at authorisation time and expires after an hour, with refresh support. Call limits apply per organisation across every authenticated client: 30/hour and 100/day on Developer, 100/hour and 500/day on Pro, and 300/hour and 5,000/day on Enterprise. Clients without native remote MCP support connect through mcp-remote.

Related MCP Servers

MCP: Checkly MCP Server

by Checkly

Checkly's hosted MCP server gives an agent read access to your synthetic monitoring account — check health, failing results, test sessions, Rocky AI root-cause analyses, incidents and status pages — over streamable HTTP with OAuth.

Cloud & Infrastructure

MCP: ESP-IDF Tools MCP Server

by Espressif Systems

Espressif's local MCP server built into idf.py — lets an AI client set the chip target, build, flash and clean an ESP-IDF project and read back config, build status and connected serial ports.

Developer Tools

MCP: Google Home MCP

by Google

Featured

Google's official MCP server for the Home ecosystem — let an agent list your homes and devices, read live trait state, run actions and query event history over OAuth.

Productivity
Browse all MCP servers →