mcp server

ongrid

An ops AI Agent that understands your infrastructure, finds the root cause, and fixes it — right from Slack, Telegram, Lark or DingTalk.

ongridio/ongridcommunitylistedunreviewednot running here yet

Use tool

Set up

Runs on your computer. Pick your agent and operating system.

This project does not publish an install method we can turn into a config. Follow the README instead.

README

synced Oct 14, 2026

Ongrid

An ops AI Agent that understands your infrastructure, finds the root cause, and fixes it — right from Slack or Telegram.

Metrics · logs · traces · topology blast-radius · root-cause correlation · remote execution · alert-driven auto-investigation · RAG knowledge & code search · specialist agents & skills.

Features • Install • Integrations • License

Features

🤖 Coordinator + Specialist agents — coordinator dispatches to SRE / network / DB sub-agents
🚨 Auto-investigate on alert — investigator spawns an RCA worker, writes the cause back to chat
🔍 Root-cause RCA — walks topology, correlates m/l/t, pins the "why" to a source-code line
🔒 Zero inbound ports — edge dials out; no port 22 / 80 / 443 on hosts
💻 Browser SSH — reverse-tunnel shell into any host; no keys, no jumpbox, all audited
🐳 Self-host in one command — install.sh brings up the full stack
📊 Built-in observability — Prometheus + Loki + Tempo + Grafana wired; the agent writes the queries
🧠 Bring your own model — Anthropic / OpenAI / GLM / DeepSeek / Gemini / Kimi, hot routing
💬 Two-way IM channels — Slack / Telegram / Larksuite / DingTalk / WeCom, per-channel locale
🛠️ Read-only host tools — bash sandbox + 26+ inspection tools; every call audited
☸️ Kubernetes lifecycle — enroll clusters, inspect workloads and events, manage upgrades, and mirror Kubernetes resources into topology
🌐 Network device management — discover neighbors from Edge hosts, verify with SNMP, poll interfaces, and map host-to-network-device links

Install

Download the latest release for your server architecture (linux-amd64 or linux-arm64), extract it, and run the installer (Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9):

Choose the command for your server architecture:

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.18.1/ongrid-v0.18.1-linux-amd64.tar.xz
tar -xf ongrid-v0.18.1-linux-amd64.tar.xz && cd ongrid-v0.18.1-linux-amd64
sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.18.1/ongrid-v0.18.1-linux-arm64.tar.xz
tar -xf ongrid-v0.18.1-linux-arm64.tar.xz && cd ongrid-v0.18.1-linux-arm64
sudo ./install.sh

🇨🇳 Mainland China — if GitHub is slow, use the matching CDN mirror URL instead:

# AMD64
wget https://ongrid.cloud/dl/ongrid-v0.18.1-linux-amd64.tar.xz

# ARM64
wget https://ongrid.cloud/dl/ongrid-v0.18.1-linux-arm64.tar.xz

Product Tour

Root Cause Analysis

Start from an operator question or alert, collect topology and device context, and produce an evidence-backed analysis with concrete next steps.

Workflow Builder

Wire triggers, agents, tools, conditions, and notifications into repeatable automations that stay editable and reviewable.

Skills Catalog

Show the tools agents can call, with clear descriptions and a visible inventory for operators.

MCP Servers

Register external MCP servers and expose their tools to chat agents and workflows with clear runtime and trust boundaries.

Knowledge Vault

Index runbooks, notes, incident history, and repositories so both humans and agents can search the same operational context.

Artifacts Center

Keep generated pages and reports in one place, with private-by-default sharing and a clean handoff for review.

Monitoring

Inspect fleet health, logs, traces, and alert state from the same workspace where the agent gathers evidence.

Topology Map

Visualize dependencies and blast radius so incidents can be traced through the affected service graph.

Approval and Write Gate

Keep risky actions behind approvals, with a visible policy boundary before anything changes production systems.

Documentation

The full product documentation is available at ongrid.cloud.

Area	Start here
Get started	Introduction · Quickstart · Architecture · Concepts
Install & operate	Server install · Edge install · First boot · Upgrade
Capabilities	Alerts · RCA · Monitoring · Logs · Traces · Knowledge · Skills
Agents	Overview · Coordinator · Incident investigator · Specialists · Reviewer
Reference	API · CLI · Alert rules · Skill manifest · Data plane

Integrations

Drop-in for the observability, channel, and model stacks your team already uses.


Observability
Channels
Models

License

AGPLv3 — see LICENSE.

Ongrid brand assets are not licensed under AGPLv3. See TRADEMARK.md.