Get startedStart · Local
Run Score Studio locally
Run the Apache 2.0 software on infrastructure you operate. You supply storage and inference; help comes through community docs and discussions.
Source reviewed 2026-09-16. Availability depends on your installation, permissions, and compatible runtimes. A supported path is not a guarantee of model quality or production readiness.
Outcome
Start a private Community installation and understand what you operate yourself.
Start here
What you need to know first
Check the inputs below before starting. If you are new, begin with the first-project guide. A dataset holds media and labels; a model produces results; a deployment makes a selected model version callable. Creating one does not create the others.
Bring these inputs
- — Docker Engine with Compose v2, 16 GB RAM, and 20 GB of free disk space
- — A reviewed Score Studio release and provider credentials for any external execution
Expected output and limits
- — A localhost-only Score Studio installation with persistent database and object volumes
- — A Community workspace whose APIs, agents, and provider connections become available when you configure and operate them
Core workflow
Start the local stack
Clone the repository and run docker compose -f docker-compose.community.yml up --build -d. The reference stack runs on your machine and binds to localhost by default. Its PostgreSQL, Redis, and MinIO services run as part of your installation.
Create the first account
Open http://localhost:3000 and register. Community runs locally with the capabilities included in the published release. It is separate from Score Cloud, and its capacity depends on the infrastructure you deploy. The taskbar identifies the installation as Local.
Choose local or connected execution
Supply compatible object storage and connect an inference or compute runtime that you operate or pay for. The Community image contains no starter model weights and has no Score-managed Fast or Ultra capacity.
Operate your installation
Your team operates the infrastructure, providers, backups, upgrades, security, and uptime. Documentation, help, and discussion are maintained with the open-source community; paid Score response times do not apply.
What this surface supports
Apache 2.0 source and local API
Storage, inference, and model weights supplied locally
Local provider and agent connections you deploy
Community-maintained docs and discussions
Expert section
Contracts, signals, and failure modes
Use this section when you are defining acceptance criteria, automating the surface, or reviewing whether its output is safe to promote downstream.
Quality and operating signals
- — All Compose services are healthy
- — The API readiness check returns ready
- — The taskbar identifies the installation as Local
- — Disk, memory, job, backup, and provider usage stay within operator limits
Common failure modes
- — Exposing the development stack before replacing keys and adding TLS
- — Deleting named volumes without a verified backup
- — Assuming Fast or Ultra capacity is bundled with Community
- — Running models without enough host memory, disk, or a compatible local runtime
Expert release checklist
□ Inputs and dependencies are pinned to immutable versions.
□ Acceptance metrics include critical classes and operating slices.
□ Failure, retry, cost, and rollback behavior are understood.
□ The resulting artifact has an owner and a downstream review path.
Engineering safeguards
- — Loopback-only default ports
- — Organization-scoped records and credentials
- — Operator-owned backups, keys, updates, and network security