Deployment

Deploy Once. Run Everywhere.

Web widget, REST API, WhatsApp, Slack, voice, email — deploy your agents to any channel with one click. Blue/green rollouts and on-premise options included.

Deploy Your First AgentView Deployment Docs
Channels

One Agent, Every Channel

Build your agent once. Deploy it to web, mobile, messaging platforms, voice, and email — all from the same configuration.

available

Web Widget

Embeddable chat widget for any website

available

REST API

Standard HTTP API with streaming

available

WhatsApp

WhatsApp Business API integration

available

Slack

Slack bot with Block Kit support

available

Voice (Twilio)

Phone and IVR integration

available

Email

Inbound/outbound email processing

beta

Mobile SDK

Native iOS and Android SDKs

beta

Telegram

Telegram Bot API integration

soon

Instagram DM

Instagram messaging integration

One-Click Deploy

From Studio to Production in Seconds

No CI/CD pipelines to configure, no infrastructure to provision. Build, test, deploy — all from the Studio.

1

Build in Studio

Design your agent in the visual canvas — prompts, tools, memory, guardrails. Test with the built-in simulator.

2

Configure Channel

Select your deployment channel. Enter API keys for WhatsApp/Slack, or copy the widget embed code.

3

Deploy

One click deploys your agent to production. Rollback is instant if anything goes wrong.

Widget Embed Code

<!-- Add to your website's <body> tag -->
<script
  src="https://cdn.orchstack.com/widget.js"
  data-agent-id="ag_cora_29xk4"
  data-theme="auto"
  data-position="bottom-right"
  async
></script>

One script tag. Fully customizable. Auto-updates when you redeploy.

Zero-Downtime Updates

Blue/Green Deployments

Ship agent updates without disrupting live users. Canary testing, instant rollback, and version pinning included.

Blue (Current)

v2.3.1
Serving 95% of traffic
Stable for 14 days
0.1% error rate

Green (New)

v2.4.0
Canary: 5% of traffic
Health checks passing
Latency within bounds
Traffic will shift to Green automatically once canary metrics are stable for 15 minutes

Zero Downtime

New version runs alongside the old. Traffic switches only after health checks pass. Users never see an error.

Instant Rollback

If the new version has issues, traffic routes back to the previous version in under 2 seconds. No manual intervention.

Canary Testing

Route a percentage of traffic to the new version first. Monitor metrics before committing to a full rollout.

Version Pinning

Pin specific tenants or channels to specific agent versions. Critical customers stay on stable while others get updates.

Deployment Models

Choose Your Deployment Model

From fully managed SaaS to completely self-hosted. Pick the deployment model that matches your security and compliance requirements.

Shared Platform

Multi-tenant SaaS deployment. Fastest time to production with automatic updates, scaling, and maintenance handled by OrchStack.

  • Instant provisioning
  • Auto-scaling
  • Managed updates
  • 99.9% SLA

Dedicated Instance

Private deployment on cloud infrastructure. Your own OrchStack instance with dedicated compute, storage, and network isolation.

  • Dedicated resources
  • Network isolation
  • Custom domains
  • Priority support

On-Premise

Run OrchStack on your own infrastructure. Full data sovereignty with Docker Compose or Kubernetes deployment options.

  • Full data sovereignty
  • Air-gapped support
  • Custom LLMs
  • No external calls
Self-Hosted

On-Premise for Enterprise

Run the full OrchStack platform in your own infrastructure. Docker Compose for simplicity, Kubernetes for scale, air-gapped for security.

Docker Compose

Single docker compose up brings up the entire platform — API server, workers, database, Redis, and vector store. Production-ready in minutes.

Kubernetes Helm Charts

Official Helm charts with HPA, pod disruption budgets, rolling updates, and health probes. Scales from 3-node clusters to hundreds.

Air-Gapped Support

Run OrchStack without internet access. Offline LLM support via local models (Ollama, vLLM). No telemetry, no external calls.

Self-Hosted Quick Start

# Clone the self-hosted repo
$ git clone https://github.com/orchstack/self-hosted.git

# Configure environment
$ cp .env.example .env
$ nano .env  # Set your LLM API keys

# Start everything
$ docker compose up -d

# Verify
$ curl http://localhost:3000/api/health
{"status":"ok","version":"2.4.0"}

Full platform running in under 5 minutes. No external dependencies required.

All Channels

Deploy to Any Channel

Build once, deploy everywhere. Every channel adapter handles format translation, media, and platform-specific features automatically.

Widget Embed

Drop a single script tag into any website to deploy a fully-branded chat widget. Customizable position, theme, and behavior with zero code changes.

API Endpoint

RESTful API with streaming support, session management, and webhook callbacks. Integrate agents into any application with standard HTTP calls.

WhatsApp

Native WhatsApp Business API integration. Rich messages, quick replies, media handling, and session continuity across conversations.

Slack

Deploy agents as Slack bots. Direct messages, channel mentions, slash commands, and interactive Block Kit responses supported natively.

Voice

Voice-enabled agents via Twilio, Vonage, or custom SIP. Speech-to-text, text-to-speech, and real-time conversation management built in.

On-Premise

Self-hosted deployment with Docker Compose or Kubernetes Helm charts. Full platform runs in your infrastructure with zero external dependencies.

Deployment FAQ

Deploy Agents Everywhere

From web widgets to WhatsApp to on-premise — your agents go where your users are.

Start DeployingRequest Demo

Free tier includes all channels -- self-host available