wabe Disk documentation

wabe Disk exposes named disks over HTTP. Each disk holds files, key/value namespaces, Redis data and named SQLite databases. Requests can read and write data, execute SQL, run a shell pipeline or execute JavaScript against that disk.

Start with a request

Follow the quickstart to write a file, execute a command and store a JSON value. Use a project API key from Dashboard. To start a server yourself, use Run wabe Disk.

Features

Feature	Quickstart
Filesystem	Write and read a file over HTTP.
Redis	Set and read a key in a Redis namespace.
SQLite	Create a table, insert a row and query it.
Actors	Register an HTTPS endpoint and call a named instance.
Git	Stage a file and commit it to a branch.

Choose an interface

Interface	Use it for
HTTP API	Read and write files, execute scripts and manage disk persistence from another service.
Shell	Search files and compose text operations with pipes and redirects.
JavaScript	Combine file operations and application state in one invocation.
Key/value storage	Store typed values, scan a key prefix and publish a batch of changes.

SQLite is the default SQL engine. The experimental Turso HTTP API requires a self-hosted build; it uses separate named databases. Shell and JavaScript wabesql calls continue to use SQLite.

Read the contracts

Durability explains when a successful write is persisted. Streaming describes file bodies and execution records. Limits lists input and runtime bounds. Errors covers failed requests and retry decisions.

Examples use the hosted wabe Cloud API. See self-hosting for standalone configuration. Self-hosted JavaScript execution requires Linux. The shell implements a documented command subset, including Git operations on filesystem branches. SQL databases are shared across those branches. A mounted POSIX filesystem and per-user permissions are outside the current API.