MSPs / MSSPs
Multi-tenant fleets under one license and org model.
CloFix Infotech · DevOps Manager
CloFix One
Your server, managed by chat.
Diagnose → Approve → Fix → Verify - with human control on every write. MSP-grade DevOps & Autonomous AI SRE to monitor, secure, troubleshoot, and operate your fleet from one place.
Ask SRE - your server, managed by chat
01Chat
02Diagnose
03Analyze
04Propose
05Approve
06Execute
07Verify
08Postmortem
What is CloFix One?
A full DevOps control plane with an Autonomous AI SRE at the center. Operators talk to servers in natural language. CloFix One gathers real host evidence through an outbound agent, explains the root cause, proposes a safe fix plan, waits for human approval, executes only allowlisted actions, and verifies the result - then drafts a postmortem.
No blind automation. No invented host state. No browser SSH for AI operations. AI proposes. Humans approve. Agents execute - only what’s allowed.
Who it’s for
Organization → Client → Environment → Server → Service tenancy for multi-tenant and enterprise fleets.
DevOps / SRE teams
Chat-first incident response with full audit trails.
IT operations
Servers, websites, SSL, SSH, security, and tickets in one UI.
Platform owners
Agents, dashboards, AI, GitOps, and licensing together.
Core pillars
Autonomous AI SRE with human approval - plus a full DevOps platform in one pane of glass.
Chat-first operations
Natural-language SRE. Select a server, describe the symptom, get evidence-backed answers and fix plans - with slash commands, @ mentions, and inline cards.
Troubleshoot & Fix
Playbook-driven investigation: probes → hypotheses → remediation → risk stamp → Approve → precheck → execute → verify desired state.
SRE Story
Per-conversation memory of environment, stack, and desired state. Long setup asks become structured Story → Plan workflows.
Safe state change
Allowlisted start/stop/restart/reload with mutex awareness (e.g. stop Apache before Nginx on port 80) - still requires Approve.
Knowledge & config library
Org knowledge base and approved config patterns inform answers. Knowledge never executes on hosts.
Outbound Go agent
Lightweight agent per host: metrics, heartbeat, allowlisted shell, files, Docker, K8s read, packages, config apply. Hosts connect out - no inbound SSH for AI SRE.
Autonomous AI SRE
Fix incidents in conversation - with proof, not guesses.
| Feature | Capability |
|---|---|
| Incident diagnose | Real probes via online agent - metrics, logs, services, ports, failed units, HTTP local checks |
| Root-cause analysis | Structured hypothesis ranking with confidence and cited evidence IDs |
| Fix plans | Allowlisted steps only; risk from read-only to critical |
| Human Approve | Every write waits on Approve / Reject (or type fix) |
| Execute & verify | Prechecks, live step output, verify active/desired state, postmortem draft |
| One run per server | No overlapping diagnose/execute chaos on the same host |
| Guardrails | Jailbreak / destructive-intent guard before planning |
Full DevOps platform
Fleet + AI SRE + Security Hub + VAPT + GitOps in one product.
Infrastructure
Clients, environments, servers, services, file manager, web terminal + audit.
Observability
Thresholds, alerts, logs, website uptime, domains, SSL expiry (30/14/7/0).
Automation
Scripts, jobs, workflows, config templates, and builds.
Deploy & cloud
Deployments, GitOps, Docker, Kubernetes, cloud accounts.
SSH center
Sessions, attacks, auth, users, keys, policies, alerts, audit.
Security & VAPT
Security Hub presets plus VAPT Lab with PDF/JSON/CSV reports.
ITSM
Incidents, RCA, backups, Kanban tasks, reports, operations wallboard.
Secrets
Vault, credential vault, temporary password share.
Admin & tools
Users, RBAC, audit logs, licensing, plus Tools Hub utilities.
Safety & governance
Reads are live. Writes wait for you.
| Control | Behavior |
|---|---|
| Human Approve for writes | Restarts, stops, config edits, package installs, workflows, generated scripts |
| Reads run live | Status, metrics, logs, and diagnostics do not need Approve |
| Dual allowlists | API and Go agent both reject forbidden commands and units |
| Risk stamping | Read-only → low → medium → high → critical; high/critical always Approve |
| Config six gates | Path allowlist · size/diff limits · syntax check · human diff · backup+atomic write · verify + rollback |
| Knowledge never executes | RAG / library / articles inform planning only |
| Hash-chained audit | AI runs, terminal, config apply/rollback, and platform actions |
| Online agent required | Offline hosts are not “faked healthy” or silently fixed |
How it works
Dashboard never SSH’s into hosts for AI SRE. Agents enroll and keep a persistent outbound connection.
Browser · CloFix One dashboard
▼
CloFix API · MySQL / Redis / Workers
▲ outbound WebSocket
CloFix Agent · on each server
License server signs plans: Free → Person → Standard → Team → Enterprise / MSSP.
All lifecycle steps
Chat → Diagnose → Analyze → Propose → Approve → Execute → Verify → Postmortem - one controlled run.
STEP 01
Chat
Describe the symptom in natural language inside Ask SRE.
STEP 02
Diagnose
Outbound agent runs live probes: metrics, logs, services, ports, failed units.
STEP 03
Analyze
Rank hypotheses with confidence and cited evidence IDs.
STEP 04
Propose
Build an allowlisted fix plan with risk stamped from read-only to critical.
STEP 05
Approve
Human reviews the plan - Approve or Reject before any write.
STEP 06
Execute
Prechecks run, then only approved allowlisted actions execute on the host.
STEP 07
Verify
Confirm active/desired state - not just exit code 0.
STEP 08
Postmortem
Draft a postmortem and keep knowledge for the next incident.
Why CloFix One
vs chatbots
Talks to real hosts via agent; cites probe evidence.
vs raw SSH scripts
Approval gates, mutex awareness, config rollback, audit chain.
vs ticketing-only ITSM
Diagnose and remediate in the same run.
vs single-purpose APM
Fleet + AI SRE + Security Hub + VAPT + GitOps in one product.
Operate your fleet by chat
Every write is allowlisted, risk-rated, and audited. Request a demo of CloFix One Ask SRE.