Run Dify at 55% Lower Cost
Move self-managed Dify workflows from local Docker to persistent CPU infrastructure with public endpoints, background workers, and storage. Connect the model provider you choose—or a separate self-hosted inference endpoint—while keeping the Dify application stack on CPU.
Fluence
$2.54/mo
Hetzner
$14.00/mo
DigitalOcean
$31.50/mo
AWS
$53.50/mo
Run Dify at 55% Lower Cost
Move self-managed Dify workflows from local Docker to persistent CPU infrastructure with public endpoints, background workers, and storage. Connect the model provider you choose—or a separate self-hosted inference endpoint—while keeping the Dify application stack on CPU.
Fluence
$2.54/mo
Hetzner
$14.00/mo
DigitalOcean
$31.50/mo
AWS
$53.50/mo
Built for always-on Dify hosting
Dify’s visual canvas is only one part of the stack. Give its apps, APIs, workers, schedules, databases, and vector state a long-lived home beyond your local Docker session.
Public apps and APIs
Publish Dify web apps and REST APIs through a reachable public endpoint. Configure a public trigger URL so external systems can invoke your webhook and integration workflows.
Workflows that keep working
Run scheduled and event-driven Workflow apps without depending on a laptop that sleeps, disconnects, or reboots.
Start from $9 per month
Use flat daily infrastructure pricing and zero egress fees for the Fluence compute portion of your stack.
Built for always-on Dify hosting
Dify’s visual canvas is only one part of the stack. Give its apps, APIs, workers, schedules, databases, and vector state a long-lived home beyond your local Docker session.
Public apps and APIs
Publish Dify web apps and REST APIs through a reachable public endpoint. Configure a public trigger URL so external systems can invoke your webhook and integration workflows.
Workflows that keep working
Run scheduled and event-driven Workflow apps without depending on a laptop that sleeps, disconnects, or reboots.
Start from $9 per month
Use flat daily infrastructure pricing and zero egress fees for the Fluence compute portion of your stack.
One architecture for Dify workflow and agents
Run Dify’s application, orchestration, and data services on Fluence CPU Cloud. Keep model inference outside the base VM by default, or add a separate GPU endpoint when local models require one.
Run on Fluence CPU Cloud
Dify web console and published applications REST API and WebSocket service Background workers and workflow scheduler PostgreSQL and Redis Vector database and file storage Plugin daemon, reverse proxy, and sandboxes
Connect externally
Hosted LLM, embedding, and reranking providers Custom or self-hosted model endpoints Optional Fluence GPU inference endpoint Email, CRM, search, and other business tools DNS, TLS, backup, and monitoring services
One architecture for Dify workflow and agents
Run Dify’s application, orchestration, and data services on Fluence CPU Cloud. Keep model inference outside the base VM by default, or add a separate GPU endpoint when local models require one.
Run on Fluence CPU Cloud
Dify web console and published applications REST API and WebSocket service Background workers and workflow scheduler PostgreSQL and Redis Vector database and file storage Plugin daemon, reverse proxy, and sandboxes
Connect externally
Hosted LLM, embedding, and reranking providers Custom or self-hosted model endpoints Optional Fluence GPU inference endpoint Email, CRM, search, and other business tools DNS, TLS, backup, and monitoring services
Get Dify running in 3 easy steps
Start with Dify’s official Docker Compose deployment path, then configure the public, persistent, and operational pieces a local installation does not need.
1
Provision the host
Choose a region, server type, storage, operating system, and networking settings for your Dify backend.
2
Configure and start Dify
Install Docker and Docker Compose, clone a tagged Dify release, and copy the example environment file: cd dify/docker cp .env.example .env Replace the default secrets, configure public URLs, and start the stack: docker compose up -d
3
Connect, publish, and operate
Add a model-provider key or custom model endpoint, publish your applications and triggers, and verify that every required container is healthy. Before production, configure TLS, monitoring, backups, restore tests, firewall rules, certificate renewal, and a controlled upgrade process.
Get Dify running in 3 easy steps
Start with Dify’s official Docker Compose deployment path, then configure the public, persistent, and operational pieces a local installation does not need.
1
Provision the host
Choose a region, server type, storage, operating system, and networking settings for your Dify backend.
2
Configure and start Dify
Install Docker and Docker Compose, clone a tagged Dify release, and copy the example environment file: cd dify/docker cp .env.example .env Replace the default secrets, configure public URLs, and start the stack: docker compose up -d
3
Connect, publish, and operate
Add a model-provider key or custom model endpoint, publish your applications and triggers, and verify that every required container is healthy. Before production, configure TLS, monitoring, backups, restore tests, firewall rules, certificate renewal, and a controlled upgrade process.
Save up to 55% on Dify base VM costs
Use lower-cost CPU infrastructure for the Dify runtime, then add inference, storage, and external services only where the workload requires them.
No egress fees
Transparent daily pricing
Public IPv4 and persistent storage available
Note: Calculator estimates reflect only the Virtual Server layer used for LlamaIndex hosting.
Save up to 55% on Dify base VM costs
Use lower-cost CPU infrastructure for the Dify runtime, then add inference, storage, and external services only where the workload requires them.
No egress fees
Transparent daily pricing
Public IPv4 and persistent storage available
Note: Calculator estimates reflect only the Virtual Server layer used for LlamaIndex hosting.
Dify AI on Fluence vs. a local machine
Local Docker is ideal for development and evaluation. Fluence is the stronger fit when Dify needs a public, persistent, and remotely accessible home.
Capability
Dify on Fluence
Dify on a local machine
Public access
Public IPv4 with customer-configured domain and TLS
Requires tunneling, port forwarding, or local network configuration
Webhooks and schedules
Continue while the VM and Dify services remain healthy
Stop when the device or Docker stack stops
Persistent data
VM storage plus optional persistent volumes and customer-managed backups
Stored on a device-bound local disk
Team access
Shared remote deployment for authorized users
Tied to one workstation or local network
Scaling path
Move to a larger configuration or split services as requirements grow
Limited by local hardware
Dify AI on Fluence vs. a local machine
Local Docker is ideal for development and evaluation. Fluence is the stronger fit when Dify needs a public, persistent, and remotely accessible home.
Dify on Fluence
Dify on a local machine
Public access
Public IPv4 with customer-configured domain and TLS
Requires tunneling, port forwarding, or local network configuration
Webhooks and schedules
Continue while the VM and Dify services remain healthy
Stop when the device or Docker stack stops
Persistent data
VM storage plus optional persistent volumes and customer-managed backups
Stored on a device-bound local disk
Team access
Shared remote deployment for authorized users
Tied to one workstation or local network
Scaling path
Move to a larger configuration or split services as requirements grow
Limited by local hardware
Why run Dify on Fluence
Persistent CPU for the full stack
Run Dify’s web app, APIs, workers, scheduler, PostgreSQL, Redis, vector database, plugins, and supporting services on a long-lived Linux VM.
Public IP and storage control
Assign public IPv4 addresses and attach persistent storage through the Fluence Console or Cloud API.
Separate runtime from inference
Keep Dify’s orchestration layer on CPU. Connect hosted model providers, a custom inference service, or an optional GPU endpoint only when the model layer requires it.
Predictable base infrastructure
Use transparent daily pricing and no Fluence egress fees while tracking model usage, storage, databases, monitoring, and external services as separate cost layers.
Why run Dify on Fluence
Persistent CPU for the full stack
Run Dify’s web app, APIs, workers, scheduler, PostgreSQL, Redis, vector database, plugins, and supporting services on a long-lived Linux VM.
Public IP and storage control
Assign public IPv4 addresses and attach persistent storage through the Fluence Console or Cloud API.
Separate runtime from inference
Keep Dify’s orchestration layer on CPU. Connect hosted model providers, a custom inference service, or an optional GPU endpoint only when the model layer requires it.
Predictable base infrastructure
Use transparent daily pricing and no Fluence egress fees while tracking model usage, storage, databases, monitoring, and external services as separate cost layers.
What you can build with Dify on Fluence
Knowledge-grounded customer support
Build a support assistant that retrieves answers from product documentation, policies, and internal knowledge using embeddings and a vector database.
Intent-based email routing
Classify incoming support messages and send each one to the correct queue, team, or downstream workflow while keeping the routing logic visible and auditable.
Scheduled reports and notifications
Run recurring workflows that gather data, generate summaries, and send daily reports, briefings, alerts, or notifications on a timetable.
Agentic RAG and research
Give an agent access to retrieval and tools so it can plan, search, retry, use fallbacks, and synthesize results across multi-step questions.
What you can build with Dify on Fluence
Knowledge-grounded customer support
Build a support assistant that retrieves answers from product documentation, policies, and internal knowledge using embeddings and a vector database.
Intent-based email routing
Classify incoming support messages and send each one to the correct queue, team, or downstream workflow while keeping the routing logic visible and auditable.
Scheduled reports and notifications
Run recurring workflows that gather data, generate summaries, and send daily reports, briefings, alerts, or notifications on a timetable.
Agentic RAG and research
Give an agent access to retrieval and tools so it can plan, search, retry, use fallbacks, and synthesize results across multi-step questions.
A global marketplace of compute
Compare available providers and locations, provision the CPU instance for your Dify runtime, attach storage and public IPs, and add GPU resources only when self-hosted inference requires them.
A global marketplace of compute
Compare available providers and locations, provision the CPU instance for your Dify runtime, attach storage and public IPs, and add GPU resources only when self-hosted inference requires them.
FAQ
Can I self-host Dify on Fluence?
How much CPU and RAM does Dify need?
Does Dify need a GPU?
Which model providers can I connect to Dify?
Can Dify expose APIs, webhooks, and scheduled workflows?
Show more
FAQ
Can I self-host Dify on Fluence?
How much CPU and RAM does Dify need?
Does Dify need a GPU?
Which model providers can I connect to Dify?
Can Dify expose APIs, webhooks, and scheduled workflows?
Show more
Be the first to know
Enter your email
Be the first to know
Enter your email