Hosted in Germany • GDPR-ready

Run AnythingLLM in Docker. Keep your data private.

Deploy a private RAG assistant in 10 minutes. All your documents stay on your infrastructure. No API calls to third parties. No vendor lock-in.

CCRMAAnalyticsAAutomationBBlogFForms
26K
GitHub stars
👥
100+
Contributors
10 min
Setup time
🎁
FREE
Open source

Deploy AnythingLLM Docker in 4 Steps

Docker Compose handles everything. No Kubernetes. No complex orchestration. Just a single file that pulls the image, sets up the database, and runs your RAG.

console.opsily.com/deploy
1
App
2
Region
3
Plan
4
Domain

Choose Your App

Select an app to get started.

1

Pull the Docker image

docker pull mintplexlabs/anythingllm:latest pulls the official AnythingLLM container. It includes Node.js, Python, and all RAG dependencies. About 1.2GB download.

2

Configure environment variables

Set your LLM provider (OpenAI, Claude, Ollama). Point to your vector database (Milvus, Weaviate, or use built-in SQLite). Specify storage paths. All in a .env file.

3

Run docker-compose up

Docker Compose spins up AnythingLLM, PostgreSQL for metadata, and optional Milvus for vector embeddings. Mounts volumes so documents persist across restarts.

4

Open localhost:3001 and start chatting

Upload documents. Select your LLM. Start asking questions. Your data never leaves your server. No analytics tracking. No data residency questions.

The Docker Advantage

Why Docker for AnythingLLM?

Docker containers are portable. Move your entire AnythingLLM stack across servers without reinstalling dependencies. Scale to multiple replicas behind a load balancer. Stop and start cleanly without orphaned processes.

GPU acceleration works. Mount your NVIDIA GPU or AMD ROCm device into the container. AnythingLLM detects it automatically and offloads inference. 5-10x faster responses for local LLM models.

Persistence is simple. Define volumes in docker-compose.yml. Your documents, chat history, and workspace config survive container restarts, updates, and crashes. Named volumes give you managed storage. Bind mounts let you see files locally.

This is how 10,000+ teams run AnythingLLM. And it's why some of them eventually switch to Opsily: after 6 months of managing Docker, updates, backups, and monitoring, the EUR 20-100/month becomes a bargain.

Why AnythingLLM Docker Teams Switch to Opsily Managed

Docker is free. Your time is not. Here's what Opsily handles so you don't have to.

No DevOps overhead

Forget monitoring uptime, managing disk space, patching security vulnerabilities, and debugging why your container won't restart. Opsily runs health checks 24/7, scales storage automatically, and applies security updates overnight. You open your laptop and it just works. That's worth the price alone.

Automatic updates

New AnythingLLM version drops. Opsily tests it, rolls it out, and keeps your documents safe. Docker forces you to read release notes, test locally, and schedule downtime. We do that for you. Zero downtime. Chat history preserved. LLM configuration intact.

Encrypted backups & compliance

Daily encrypted snapshots. Stored separately from your live data. We handle GDPR data residency in EU servers, HIPAA audit logs, SOC2 compliance. Docker is just a container; compliance audits ask who backs up what and where it lives. Opsily answers those questions for you.

Built for teams who need reliability

99.9%
Uptime SLA
24/7
Monitoring
Daily
Backups
5 min
Deploy time
Monthly Cost Breakdown
Zapier Pro$29.00
HubSpot Starter$45.00
Typeform Basic$25.00
Total SaaS Cost$99.00/mo
Opsily Server
$20.00/mo
You save $948/year
EUR 1.2K

Saved per year. Per team member. After switching to managed.

Average: 3-5 hours per week managing Docker. At EUR 50/hour loaded cost. Opsily managed: EUR 20-100/month. Math favors the managed side after month 3.

See our pricing

AnythingLLM + Best-In-Class Ecosystem

AnythingLLM is the RAG UI. Pair it with these tools running in the same Docker network.

Ollama runs local LLMs like Llama 2, Mistral, or Dolphin on your hardware. No API costs. No external dependencies. Full offline support.

PostgreSQL stores your document metadata and workspace configuration. Included in the official docker-compose.yml. Handles millions of documents without breaking a sweat.

Milvus or Weaviate scale vector storage beyond SQLite. If you're ingesting 50,000+ documents, a dedicated vector DB is worth it. Both support Docker Compose.

Opsily hosts all of these pre-configured and scaled. You pick the combination that fits your workload.

Choose Your Path

Run AnythingLLM for free on your server. Or let Opsily handle infrastructure, backups, and updates. Both paths work. One path is less stressful.

Monthly
Annual

Loading pricing...

Security & Compliance Built In

Whether you self-host or use Opsily managed, your data is protected.

GDPR compliant

EU data residency on request. No data shared with third parties. Full audit trails for document access.

Encrypted backups

Daily snapshots. AES-256 encryption at rest. Stored in geographically redundant infrastructure.

No vendor lock-in

Export your documents anytime. AnythingLLM format is standard. Switch to Opsily or migrate elsewhere without data loss.

99.9% uptime SLA

Opsily managed only. Self-hosted Docker is best-effort. We guarantee availability.

Zero tracking

AnythingLLM doesn't phone home. Chat history stays on your infrastructure. No analytics, no fingerprinting.

Ready to deploy AnythingLLM?

Choose your own adventure: free Docker on your server, or Opsily managed with backups and updates handled.