Must have
Installation requirements
- The default MongoDB image needs a CPU with AVX instructions; it crashes at startup on CPUs without AVX.
Application hosting
Hosting LibreChat
LibreChat is an open-source chat interface that connects to many AI providers and to local models. It sends prompts to those providers rather than running models itself, so a small server without a GPU runs it.
Requirements last checked against the project's documentation.
A sensible first server for LibreChat is 1 vCPU, 2 GB RAM and 20 GB of storage. The cheapest listed plan that fits is OVHcloud VPS-1 (US) at $4.54/month.
- Min vCPU
- 1
- Min memory
- 1 GiB
Understand the setup
Find the software and services needed to run the application.
See requirementsBudget your resources
Plan memory, storage and CPU for the way you will use it.
Open resource plannerBuild your shortlist
Compare plans for the starting size, then review the provider and billing terms.
See plansRecommended
Planning guidance
- Start with 1 vCPU, 2 GB RAM and 20 GB storage.Resize after watching real memory and CPU use under your own traffic.
- Use the resource planner to size for your user count.
Minimum installation resources
The default stack runs LibreChat with MongoDB, Meilisearch and a RAG API backed by PostgreSQL with pgvector. Model inference runs at the providers you connect.
Where to start
A sensible first server for LibreChat
- vCPU
- 1
- RAM
- 2 GB
- Disk
- 20 GB
LibreChat's Docker guide asks for at least 1 vCPU and 1 GiB of RAM for its containers and 2 GB for smooth operation with every service running. Disk is our estimate; uploaded files and the search index grow with use.
Our shortlist
Cheapest plans that meet this size
-
$4.54/mo list price
-
$5.00/mo list price
-
$9.99/mo list price
-
$10.00/mo list price
-
$10.00/mo list price
-
$10.00/mo list price
-
$10.00/mo list price
-
$10.00/mo list price
One cheapest plan per provider, based on each provider's standard list price. Check the plan page for introductory and renewal pricing before you buy.
Deploy with
Supported paths
The host must provide
Runtime
- MongoDB
- Stores users and conversations; included in the default Compose stack
- Meilisearch
- Conversation search; included in the default Compose stack
- RAG API with PostgreSQL and pgvector
- File search; included in the default Compose stack
Storage
Persistent data
Persist the MongoDB data, the Meilisearch index, the pgvector database and uploaded files. Conversations and attachments grow with every user, so plan backups for all three data stores.
Step by step
Plan a LibreChat deployment
-
Create the environment file
Copy .env.example to .env and replace every sample secret, including CREDS_KEY, CREDS_IV and the JWT secrets, with unique generated values that stay fixed across restarts.
Official instructions -
Start the Compose stack
Run the default Docker Compose file, which starts LibreChat with MongoDB, Meilisearch and the RAG API. LibreChat listens on port 3080 unless PORT is set.
Official instructions -
Add providers and custom settings
Add provider API keys, and mount librechat.yaml through docker-compose.override.yml for custom endpoints and settings.
Official instructions
Security
Before you go public
- Generate unique CREDS_KEY, CREDS_IV and JWT secrets; LibreChat refuses to start with retired published defaults.
- Serve it over HTTPS through a reverse proxy and keep MongoDB, Meilisearch and PostgreSQL off the public internet.
- Decide who may register before opening the instance to others.
Cost
Beyond the server price
- The server is cheap; API usage at the providers you connect is the main running cost.
- Budget disk and backups for conversations, uploads and the vector database.
Want the full monthly bill? Use the hosting cost calculator. Add storage, backups and other costs to your estimate.