🇫🇷 FR 🇬🇧 EN
\ud83e\udd16 Muppy AI-Native Ops Offer

Your AI agent operates the infra —
every action is audited

You already use AI agents to code. Muppy extends this logic to operations: 9 MCP profiles, AI/LLM Gate with temporal granularity, and every action captured in IMQ — structured log + subprocess stdout/stderr + who/when. Zero SSH, zero ops ticket.

Request a demo → ← Back to Muppy

What the agent can do via MCP

Real actions available immediately, without intermediate scripts

\ud83e\udd16 Agent action→ Result
Provision an App ServerLXC + PG + Traefik + IDE in 10 min
Trigger a K8s deploymentHelm upgrade + DF12 journal recorded
Read upgrade logsIMQ log + subprocess console
Inspect a PG clusterReplication state, active locks, lag
List firewall rulesActive UFW + dynamic CIDRs
Create an ephemeral ruleTTL defined, automatic expiry
Consult K8s journal3-level diff + complete Helm op
Run a healthcheckTopology-aware result + alert

AI/LLM Gate — Audited access with temporal granularity

9 MCP profiles + controlled access

⏳
Temporal gate

Access validated “Until…” with until, granted_by, granted_ts. Automatic expiry.

🔐
Human approval

ai_llm_managed = False. Destructive operations → pending_approval. The human validates.

📋
Audit trail

Every operation captured in IMQ: requesting_user_id + timestamp + log + stdout/stderr.

🤖
9 MCP profiles

From read-only discovery to full write access. Profile selected per AI session context.

Ready to give your AI agents control of your infra?

Every action audited. Every operation reversible. Zero SSH.

Request a demo → See all offers