Morph Volumes
Durable S3-compatible Morph storage for build caches, release artifacts, checkpoints, and shared datasets. Volumes keep the auth model simple: reuse your Morph API key and talk to the gateway with any standard S3 client.
Auth
Morph API key
Endpoint
volumes.svc.cloud.morph.so
Pricing
$0.0054794/GiB-hr
S3-Compatible Quick Start
Path-style, SigV4, region us-east-1
export MORPH_API_KEY="<your-api-key>"
export AWS_ACCESS_KEY_ID="$MORPH_API_KEY"
export AWS_SECRET_ACCESS_KEY="$MORPH_API_KEY"
export AWS_DEFAULT_REGION="us-east-1"
aws --endpoint-url "https://volumes.svc.cloud.morph.so" s3 ls
aws --endpoint-url "https://volumes.svc.cloud.morph.so" s3 sync ./dist s3://my-volume/releases/distUse your Morph API key for both `AWS_ACCESS_KEY_ID` and `AWS_SECRET_ACCESS_KEY`. Volume names are globally unique, which makes public cache distribution predictable across teams and automation.
Persistent
Keep expensive outputs warm
Store compiler caches, package mirrors, model checkpoints, or release bundles once and share them across devboxes, CI, and agent workflows.
Simple Auth
No new credential type
Volumes use the Morph API keys you already issue today. Any standard S3 client can speak to the gateway without a custom SDK or one-off auth broker.
Usage-Based
Storage billed by GiB-hour
Current runtime pricing is $0.0054794/GiB-hr (~$4.00/GiB-month). The Usage page surfaces volume billing snapshots alongside Morph VM consumption.
Operational Notes
The volumes data plane is public and S3-compatible today. Provisioning and policy changes still happen through the Morph control plane, which keeps bucket lifecycle orchestration away from browsers for now.
Path-style access is required, the region is fixed to us-east-1, and anonymous reads only work for volumes already marked `public_read`.
This makes Volumes a good fit for durable build artifacts, remote package caches, shared research datasets, and distribution-friendly public cache layers.
What should I store in Morph Volumes?
Use Volumes for outputs that are expensive to rebuild but easy to consume from S3: compiler caches, wheelhouses, model checkpoints, test fixtures, release bundles, and durable team artifacts.