How to Deploy LibreChat on a VPS

copyapt update && apt upgrade -y
adduser deploy
usermod -aG sudo deploy

copyufw allow OpenSSH
ufw allow 80
ufw allow 443
ufw enable

copycurl -fsSL https://get.docker.com | sh
usermod -aG docker deploy

#Install LibreChat

Create a working directory:

copymkdir ~/librechat && cd ~/librechat

Create a .env file with your configuration:

copycat > .env <<EOF
# Domain and security
ALLOW_REGISTRATION=true
ALLOW_SOCIAL_LOGIN=false
ALLOW_SOCIAL_REGISTRATION=false
APP_TITLE=My AI Chat
HOST_DOMAIN=ai.example.com
HOST_PORT=443

# MongoDB
MONGO_URI=mongodb://mongo:27017/LibreChat

# Session
CREDS_KEY=$(openssl rand -hex 32)
CREDS_IV=$(openssl rand -hex 16)
JWT_SECRET=$(openssl rand -hex 32)
JWT_REFRESH_SECRET=$(openssl rand -hex 32)

# API keys (add your own)
OPENAI_API_KEY=
ANTHROPIC_API_KEY=

# Ollama (if using)
OLLAMA_BASE_URL=http://host.docker.internal:11434
EOF

Create a compose file:

copyservices:
  librechat:
    image: ghcr.io/danny-avila/librechat:latest
    restart: unless-stopped
    ports:
      - "127.0.0.1:3080:3080"
    depends_on:
      - mongo
    env_file:
      - .env
    volumes:
      - ./images:/app/client/public/images
      - ./librechat_data:/app/backend/data
    extra_hosts:
      - "host.docker.internal:host-gateway"

  mongo:
    image: mongo:6
    restart: unless-stopped
    ports:
      - "127.0.0.1:27017:27017"
    volumes:
      - mongo_data:/data/db

volumes:
  mongo_data:
  librechat_data:

Start it:

copydocker compose up -d

LibreChat listens on port 3080. The 127.0.0.1 bind keeps it private.

#Configure model providers

LibreChat supports multiple providers. Configure them in the .env file:

OpenAI:

copyOPENAI_API_KEY=sk-your-key-here

Anthropic:

copyANTHROPIC_API_KEY=sk-ant-your-key-here

Ollama:

copyOLLAMA_BASE_URL=http://host.docker.internal:11434

Google Gemini:

copyGOOGLE_API_KEY=your-key-here

After changing the .env file, restart LibreChat:

copydocker compose up -d

#HTTPS + domain

LibreChat needs HTTPS for production use. Point a reverse proxy at 127.0.0.1:3080 and terminate HTTPS on 443.

The simplest path is Automatic HTTPS with Caddy. Point an A record for your hostname (say ai.example.com) at the server's public IP, then have Caddy reverse-proxy that hostname to 127.0.0.1:3080.

If you're using the Caddy container approach, put LibreChat and Caddy in the same compose file and proxy to the LibreChat service name:

reverse_proxy librechat:3080

#First-run setup

Load https://ai.example.com in your browser. On a fresh install, LibreChat shows an admin setup screen.

Create the admin account immediately. The setup screen is open until the first account exists — whoever registers first becomes the admin. Set your email and a strong password.

Once you're in:

1. Check the model dropdown — it should show models from your configured providers. If Ollama is connected, you'll see local models alongside cloud models.
2. Send a test message — try a message with a cloud model (OpenAI/Anthropic) and a local model (Ollama). Both should work if configured correctly.
3. Explore presets — LibreChat lets you save model configurations as presets. Create a "Local Only" preset that uses only Ollama, and a "Cloud Only" preset for API models.

#Multi-user setup

LibreChat supports multiple users with role-based access. After creating the admin account:

1. Disable open registration (Settings → Registration) once you've created all the accounts you need.
2. Create user accounts (Admin Panel → Users) for each person who needs access.
3. Set default models (Settings → General) to control which model new conversations start with.

Users get their own conversation history and settings, but they all share the same API keys and model backends. There's no per-user API key isolation — everyone uses the same keys.

#Backups

LibreChat stores conversations in MongoDB and settings in the data volume. Back up both:

copy# Backup MongoDB
docker compose exec -T mongo mongodump --archive | gzip > librechat-mongo-$(date +%F).sql.gz

# Backup LibreChat data
docker run --rm -v librechat_data:/data -v $(pwd):/backup alpine \
  tar czf /backup/librechat-data-$(date +%F).tar.gz -C /data .

Restore MongoDB from a dump:

copygunzip -c librechat-mongo-2026-08-08.sql.gz | docker compose exec -T mongo mongorestore --archive

#Upgrades

Pull the newer image and recreate:

copydocker compose pull
docker compose up -d

LibreChat runs database migrations automatically. Check the LibreChat changelog for breaking changes before a major version bump.

#Troubleshooting

Can't connect to Ollama. Verify Ollama is running (docker compose ps in the Ollama directory) and the URL is correct. If both are on the same server, http://host.docker.internal:11434 should work.

API key errors. Check that your API keys are valid and have credits. For OpenAI, verify the key at platform.openai.com. For Anthropic, check console.anthropic.com.

MongoDB connection errors. Verify MongoDB is running (docker compose ps) and the MONGO_URI is correct. Check MongoDB logs for authentication errors.

"Unauthorized" errors. Your JWT secrets may have changed. If you regenerated them, you'll need to log in again with the admin credentials.

Slow responses from Ollama. Expected without a GPU. CPU inference is 5-10x slower than GPU. Consider using a smaller model or adding a GPU.

Conversations missing after restart. LibreChat stores data in Docker volumes. If volumes were removed or recreated, conversations are lost. Back up the data volumes regularly.

#Verification + next steps

You're done when you can: load the web interface over HTTPS, sign in as admin, switch between different model providers, send messages and get responses, and create additional user accounts. The model dropdown should show available models from all configured providers.

From here, explore LibreChat's advanced features like conversation branching, custom endpoints, and plugin support. For a simpler multi-model interface, see Open WebUI. For document-based AI workflows, see AnythingLLM. A Hetzner CX22 (2 vCPU / 4 GB) handles remote-backend setups; pair with a CX42 (8 vCPU / 16 GB) for local Ollama. See Best VPS for AI & ML Workloads for the ranked picks.