Today we’re shipping Uploadista SDK 1.0 — the first stable release of the open-source engine that powers file uploads and processing in Uploadista. After months of beta iterations shaped by real production feedback, the API surface is stable, the flow engine is battle-tested, and it’s ready for you to build on.
This is the release we’ve been working toward: a single, typed, composable way to move bytes from a client to storage — and do everything that needs to happen in between.
What Uploadista SDK is
The SDK treats file handling as a pipeline problem, not an upload problem. You define a flow — a directed graph of processing steps — and the engine handles orchestration, retries, parallelism, and progress tracking. Each step has typed inputs and outputs, so if a resize step expects an image and gets a video, the compiler tells you at build time.
Everything is MIT-licensed and runs where you run it: Node, Bun, and edge runtimes like Cloudflare Workers.
Highlights of 1.0
A production-ready flow engine. The DAG processor at the core is now stable. Nodes cover the whole lifecycle — input, image optimization and resizing, utility operations (conditional routing, merge, multiplex, zip), and storage output. When step 3 of 5 fails, only step 3 reruns on retry. No orphaned files, no manual cleanup.
A new queue system for flows. The headline addition in 1.0. Flows can now be enqueued and executed reliably through a dedicated queue service, giving you durable, back-pressure-aware processing instead of fire-and-forget execution. This is what makes heavy, long-running pipelines dependable at scale.
Real-time progress over WebSockets. Upload and flow progress is streamed to clients as it happens, so your UI reflects accurate state instead of guessing.
Bring Your Own Storage. First-class data stores for AWS S3, Google Cloud Storage, Azure Blob Storage, and the local filesystem. Your files stay in your buckets — a prerequisite, not a feature, for anyone with data-residency requirements.
Pluggable KV stores. Metadata and pipeline state can live in Redis, IORedis, Cloudflare Durable Objects, Cloudflare KV, an in-memory store for development, or the filesystem. Swap backends without touching your flow logic.
TypeScript-first, built on Effect. The SDK is built with Effect-TS from the ground up for robust, typed error handling. Your pipeline definitions are fully typed end to end — and 1.0 ships with support for the latest TypeScript.
Getting started
Install the packages you need and define your first flow:
npm install @uploadista/core @uploadista/serverFrom there, compose input, processing, and storage nodes into a flow and run it locally, then deploy the same definition to production. The documentation walks through the full setup.
What’s next
1.0 is a foundation, not a finish line. We’re continuing to invest in the visual flow builder — design pipelines on a canvas with the TypeScript generated for you — along with deeper documentation and native mobile SDKs. And because the core is open source, the roadmap is shaped in the open.
If file processing at scale has ever bitten you, we’d love for you to try it. Browse the SDK on GitHub, read the docs, or sign up for Uploadista Cloud — no credit card required.
Thank you to everyone who tested the betas and filed the sharp, honest issues that got us here. This release is better because of you.