Skip to main content
Omazy CLI

One CLI.
Every Omazy automation.

200+ commands, a live terminal dashboard, and 112+ MCP tools at your fingertips. Connect OpenClaw, Claude, or any agentic tool to your Omazy workspace — locally, securely, in under 30 seconds.

Explore modes

200+ automation commands

Manage channels, bots, CRM, campaigns, and more

112+ MCP tools + ACP support

MCP + ACP for multi-agent orchestration

Local-first, enterprise-secure

Credentials never leave your machine

Omazy CLIBeta

Join the Omazy CLI Beta

Get early access to 200+ commands, TUI dashboard, and MCP server for AI agent automation.

No spam. We'll notify you when your access is ready.

You're on the list!

Check your inbox — we've sent you a confirmation email. We'll notify you when your CLI beta access is ready.

One line to connect

Static binary. No runtime dependencies. ~15MB.

$ brew install omazy/tap/omazy-cli
$ curl -fsSL https://cli.omazy.com/install.sh | sh
> irm https://cli.omazy.com/install.ps1 | iex

Works with your existing Omazy workspace. Credentials stored in your OS keychain. Auto-updates in the background.

Three runtime modes. One binary.

CLI. TUI. MCP.

Built in Go with Cobra, Bubble Tea, and mcp-go. Starts in 10ms, uses under 20MB. Adapts to how you work — script it, navigate it, or hand it to AI.

CLI Mode

Script anything. Automate everything.

200+ commands across 27 groups — channels, bots, queue, CRM, campaigns, AI agents, knowledge base, and more. Every command routes through the unified gateway at hq.omazy.com with automatic auth injection and retry.

  • 27 command groups

    Auth, channels, bots, queue, CRM, campaigns, KB, AI agents, webhooks, tasks, billing, reports — the full platform

  • Pipe-friendly output

    Auto-detects TTY vs pipe — styled tables for humans, clean JSON for scripts and jq

  • Idempotent & dry-run

    Use --dry-run to preview, --idempotent to safely retry, --batch for bulk JSON input

  • Concurrent bulk ops

    Import 10,000 customers with progress bars, resume support, and per-batch error tracking

Omazy CLI mode — list channels, view queue, deploy bots, generate reports
TUI Mode

A dashboard in your terminal.

Built with Bubble Tea and Lipgloss. 8 live-updating views — queue monitor, channel explorer, bot chat with streaming, CRM browser, task board, and real-time log stream. All powered by SSE.

Omazy TUI — interactive terminal dashboard showing queue navigation, conversation details, and agent assignment

8 interactive views

Dashboard, channels, queue, bot chat, CRM, tasks, logs, and campaign — each auto-refreshing via SSE

Keyboard-driven

Tab/Shift+Tab for panes, j/k to navigate, / for fuzzy search, ? for help overlay

Streaming bot chat

Token-by-token responses in real time — like ChatGPT, but in your terminal

SSH-friendly

First paint in under 100ms. Dark/light/auto themes. Works over remote sessions.

MCP Mode

Let AI agents operate Omazy.

112+ tools across 15 categories exposed via stdio or SSE transport. Supports both MCP and ACP protocols. Includes resources, prompt templates, and persistent context — so AI agents don't need to pass app_id on every call.

  • 112+ tools, 8 resources, prompt templates

    Full platform coverage — channels, bots, queue, CRM, campaigns, AI, billing, and more

  • Stdio + SSE transport

    Local agents use stdio (Claude Desktop, OpenClaw). Remote agents connect over SSE on any port

  • Context-aware

    Persistent scope manager tracks app, session, and last-used resources across tool calls

  • Local-first security

    Credentials stay in your keychain. AI agents never see tokens — only scoped API results

Omazy MCP Server — start server and expose 112+ tools for AI agents
Claude Desktop
You: Check my open support queue
Claude: I'll check your queue now.
→ omazy_queue_list(status="open")
→ Found 23 open conversations
You have 23 open conversations.
8 are waiting over 5 minutes.
Shall I assign the top priority ones?
AI Agent Integration

Your AI agent's new superpower

With MCP Server mode, any AI assistant becomes a full Omazy operator — reading queue state, managing bots, running campaigns, and acting on customer data. Just describe what you need in plain language.

Triage conversations

"Scan my open queue, prioritize by VIP status and wait time, and assign the top 5 to available agents."

Generate reports

"Analyze this week's CSAT scores by channel and agent, flag anyone below 4.0, and draft a summary."

Manage bots

"Check which bots are active, compare their containment rates, and disable any performing below 30%."

Run campaigns

"Send a thank-you broadcast to all customers who opened a support ticket this week."

Monitor in real time

"Watch my queue and alert me when any VIP customer waits more than 3 minutes."

Audit & compliance

"Review all escalation-tagged conversations from the past 7 days and check SLA compliance."

Which mode fits your workflow?

All three modes connect to the same workspace. Use them together or independently.

Scenario CLI TUI MCP
Scripts & automationBest
One-off commandsGoodBest
Real-time monitoringBest
AI agent automationBest
CI/CD pipelinesBestGood
Bulk operationsBestGoodGood
Team dashboards over SSHBest
Built for your workflow

From support floor to agentic automation

Omazy CLI fits wherever you work — scripted, scheduled, or agent-driven.

Support Operations

Auto-assign conversations by skill, bulk-close stale tickets, and generate shift reports. Use --watch for live queue monitoring or schedule via cron.

$ omazy queue list --watch --interval 5s

Bot & AI Agent Management

Export bot configs as JSON, deploy across 10 channel types, test with streaming chat, and monitor containment rates — all from your terminal.

$ omazy bot chat bot_123 --stream

Channel & Widget Deployment

Set up WhatsApp, email, Telegram, Slack, and 6 more channel types. Generate widget embed code, manage allowed domains, and test connectivity.

$ omazy channel add --type whatsapp --interactive

CRM & Campaign Automation

Import customers from CSV, create cohorts, track events, and launch targeted campaigns. Pipe customer data into any analytics tool via JSON output.

$ omazy crm customers import data.csv --batch-size 100
Omazy for Claude

Connect Claude to your Omazy workspace

Add 3 lines to your Claude Desktop config and Claude becomes a full Omazy operator — reading your queue, managing bots, running campaigns, all through natural conversation.

One-file setup

Add the Omazy MCP server to your claude_desktop_config.json. Claude auto-discovers all 112+ tools on first connection.

Works with Claude Desktop, Cursor & Windsurf

Any MCP-compatible Claude client connects instantly. No custom API wrappers, webhooks, or middleware needed.

Context-aware conversations

The CLI maintains persistent scope — set your app once and Claude inherits the context across every tool call in the conversation.

Stdio + SSE transport

Local agents use stdio transport. Remote or cloud-hosted Claude instances connect over SSE on any port you configure.

claude_desktop_config.json
"mcpServers": {
"omazy": {
"command": "omazy",
"args": ["mcp", "serve"]
}
}
Claude Desktop
You: Check my open support queue and assign the top priority one
Claude: I'll check your queue now.
→ omazy_queue_list(status="open")
→ Found 23 open conversations
→ omazy_queue_assign(item="1042", agent="sarah")
Done! #1042 (Sarah Ahmed, WhatsApp, VIP) is now
assigned to Sarah. She has 3 active conversations.
Full platform coverage

27 command groups. Every Omazy operation.

From auth to billing, every feature of the Omazy platform is accessible through a composable, pipe-friendly CLI command.

auth
config
app
channel
widget
user
workgroup
message
queue
bot
kb
crm
canned
campaign
autoreply
webhook
task
flow
agent
ai
prompt
notification
billing
report
tui
mcp
context
Enterprise ready

Built for teams that take security seriously

Omazy CLI is designed for enterprise adoption from day one — multi-profile configs, scoped permissions, unified gateway architecture, and defense-in-depth security.

Multi-profile configuration

Manage production, staging, and development environments separately. Switch contexts instantly with omazy config profile use.

Scoped API keys with RBAC

Admins control exactly which operations each team member or automation can perform. Role-based permissions enforced at the gateway level.

Unified gateway architecture

All 200+ commands route through a single API origin (hq.omazy.com) with consolidated auth, rate limiting, and audit logging across 6 namespaces.

Defense-in-depth security

OS keychain storage (macOS Keychain, Windows Credential Manager, Linux Secret Service) with AES-256-GCM fallback. Static binary with no CGo, SBOM generated, code-signed releases.

Air-gapped & container ready

Single static binary works on Alpine Linux, Docker, and air-gapped networks. No Node.js, Python, or Java runtime required.

Machine-readable error handling

Structured error codes (e.g., CHANNEL_PHONE_ALREADY_REGISTERED) with actionable suggestions. --debug flag for full request/response logging.

OmazyClaw

Powered by Omazy CLI + OpenClaw

Your Omazy automation agent.

OmazyClaw is a lightweight Omazy CLI mode built for OpenClaw and personal AI agents. One install, and your lobster has full access to your Omazy workspace — locally, securely, instantly.

Explore modes
OmazyClaw — Omazy CLI for OpenClaw. Local-first MCP, OpenClaw + ACP ready, 112+ MCP tools.

What you get

Make it yours. Connect your OpenClaw agent to Omazy and let it manage conversations, bots, campaigns, and CRM — all through natural language.

Always local, zero risk. Credentials stay in your OS keychain. Your agent never sees tokens — it only receives scoped API results through the local CLI.

Ready in 10 seconds. Install Omazy CLI, run omazy mcp serve, and point your OpenClaw config at it. No API keys to copy, no webhooks to set up.

OpenClaw & personal agents

Run OpenClaw, LangGraph, CrewAI, or any custom agent on your own hardware. Omazy CLI exposes 112+ MCP tools locally — your agent discovers and calls them without touching Omazy's API directly.

Self-managed environments

Everything runs on your machine. Multi-profile configs let you isolate production, staging, and dev. Credentials stay in your OS keychain — never transmitted to agent runtimes.

Multi-agent orchestration

ACP + MCP support means multiple agents can coordinate across Omazy operations. One agent triages the queue, another generates reports, a third manages bot deployments — all through the same local CLI.

Any agent, any framework

Claude Desktop, Cursor, GPT via API, AutoGen, Semantic Kernel — if it speaks MCP or ACP, it works with Omazy CLI. No vendor lock-in, no proprietary SDKs.

How it works

From trigger to action in 5 steps

See how Omazy CLI sits at the center of every automation — bridging your AI agents to the Omazy platform, locally and securely.

POST hq.omazy.com/webhook/ingest/stripe
✓ Webhook received
Event: payment_intent.succeeded
Source: Stripe (acct_1Nq...)
Customer: David Chen <david@acme.co>
Amount: $249.00 — Plan: Pro Monthly
Routing to workflow: payment_success_flow
◆ OpenClaw Agent v3.2 — Task received
[reasoning] Incoming payment event for David Chen.
[reasoning] Customer is on Pro plan. Check CRM for context.
[tool_call] omazy_crm_customers_search("david@acme.co")
[result] Found: cust_8x2f (VIP, 3 prior tickets)
[reasoning] VIP customer. Send WhatsApp confirmation + update CRM.
[tool_call] omazy_message_send(channel="whatsapp", ...)
[tool_call] omazy_crm_track(cust_8x2f, "payment_received")
$ omazy message send --channel whatsapp \
  --to "+1234567890" \
  --template order_confirmation \
  --vars '{ "name":"David","order":"#4821"}'
✓ Message sent to +1234567890 via WhatsApp
Template: order_confirmation | ID: msg_x8f2k9
# Omazy platform processes
Channel: WhatsApp Business (+1 555-012-3456)
Delivered: ✓  Read: pending
CRM: Updated David Chen interaction history
Queue: No escalation needed
# Response flows back to agent
{ "success": true, "message_id": "msg_x8f2k9",
  "delivered": true,
  "next_action": "monitor_reply" }
Agent: Task complete. Monitoring for reply...
Step 1 of 5
01

Webhook ingested

An external event arrives at hq.omazy.com/webhook/ingest from Stripe, Facebook, Segment CDP, HubSpot, or your own backend.

StripeFacebookSegmentHubSpot
02

Omazy AI or OpenClaw decides

Omazy AI agent or your OpenClaw lobster evaluates the event payload, matches it to a workflow rule, and selects which MCP tools to call.

Omazy AIOpenClawWorkflow rulesMCP tools
03

Omazy CLI executes

Running locally, the CLI authenticates via OS keychain and makes scoped API calls. No credentials exposed to agents.

112+ MCP toolsKeychain authScoped perms
04

Omazy platform acts

Operations execute on your workspace — messages, bots, campaigns, CRM updates, queue assignments.

ChannelsBotsQueueCRM
05

Results flow back

Structured JSON responses return to your agent. Chain further actions, escalate, or notify your team.

JSON responseChain actionsNotify

Every step runs on your infrastructure. Your agent never touches credentials, never calls Omazy APIs directly, and every operation is scoped to your team's permissions. This is the architecture that makes AI-driven automation enterprise-safe.

Explore all 3 runtime modes
The automation engine

Omazy CLI at the center of every workflow

Whether triggered by a webhook, scheduled by cron, or driven by an AI agent — Omazy CLI is the bridge between your automation logic and the Omazy platform.

01

Trigger

Webhook, cron job, CI/CD pipeline, or human command

02

Agent Brain

OpenClaw, Claude, LangGraph, or custom orchestrator

03

Omazy CLI

112+ MCP tools, scoped API calls, local-first auth

04

Omazy Platform

Channels, bots, queue, CRM, campaigns, knowledge base

200+

Commands

27 groups, 6 API namespaces

112

MCP Tools

15 categories + 8 resources

10ms

Startup time

Static Go binary, ~15MB

10

Channel types

WhatsApp, email, Slack & more

Frequently asked questions

Not at all. Commands use plain English like omazy queue list and omazy bot status. The TUI mode gives you a visual dashboard that needs zero command-line experience. Developers and ops teams will love the scripting and CI/CD capabilities too.

No — it's a complement. Use whichever interface fits the moment: the web dashboard for visual exploration, the CLI for speed and automation, and MCP mode for AI-driven operations. They all connect to the same workspace.

MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude and GPT call external tools. With Omazy CLI's MCP Server mode, AI agents can directly read your inbox, manage bots, run reports, and take action — no custom API code required.

Yes. MCP tools have scoped permissions — you control which operations are available. Read-only tools are separated from write tools. You can start in read-only mode and expand access as you're comfortable.

The CLI works with all Omazy plans. Some commands map to features that require specific plan tiers, but the CLI itself is free to install and use.

Omazy CLI isn't a DevOps tool — but it integrates beautifully with your CI/CD workflows. Use it in GitHub Actions, GitLab CI, or any pipeline to auto-discover and sync your widget, email, and messaging channel deployment settings with Omazy. Great for keeping your Omazy configuration in version control and deploying changes automatically when your infrastructure updates.

Yes. Omazy CLI supports ACP alongside MCP, giving your agentic tools a standardized way to communicate with the Omazy platform. ACP enables multi-agent orchestration — your AI agents can coordinate tasks, share context, and hand off work across Omazy operations. Whether you're using OpenClaw, LangGraph, CrewAI, or custom agents, ACP support means they can natively discover and invoke Omazy capabilities without custom integration code.

Omazy CLI is local-first by design. Your credentials, tokens, and configuration never leave your machine — they're stored locally in your system keychain or encrypted config files. No secrets are shared across machines or transmitted to third parties. When AI agents connect via MCP, they communicate through the CLI on your device, which makes scoped permission calls directly to the Omazy backend. This means your AI agent never touches your credentials — it only sees the results the CLI returns. All data stays local, all API calls are scoped, and you stay in full control.

Absolutely. Omazy CLI is built for enterprise-grade adoption from day one. It supports multi-profile configuration so teams can manage production, staging, and development environments separately. Authentication uses scoped API keys with role-based permissions — admins control exactly which operations each team member or automation can perform. The CLI integrates seamlessly with any tool your enterprise already uses: connect it to OpenClaw, Claude, or custom internal agents via MCP; plug it into GitHub Actions, GitLab CI, or Jenkins for deployment automation; or use it alongside Terraform, Ansible, or your existing infrastructure tooling. No vendor lock-in, no proprietary protocols — just open standards and security-first design that fits into your existing workflow without friction.

omazy cli
Omazy CLI v2.4.0 — 200+ commands • TUI dashboard • MCP server
> omazy  

Join the waitlist for Omazy CLI

Be among the first to automate your Omazy workspace from the terminal. Get early access to 200+ commands, the TUI dashboard, and MCP server for AI agent integration.

Free to use. No credit card required.

Local-first security Works on macOS, Linux, Windows OpenClaw & Claude compatible