Fuse Cloud: hosted Fuse.js search — would you use it?

[krisk]

I'm exploring Fuse Cloud, a hosted version of Fuse.js, and want to see if there's interest before going deeper.

The goal is simple: upload JSON, get a search API.

Same fuzzy matching style and Fuse-like results shape, but hosted so you don't need to ship the full dataset to the client.

import { FuseCloud } from "@fusejs/cloud"

const client = new FuseCloud({
  publicKey: "pk_abc123",
  index: "products"
})

const { results } = await client.search("iphone")

Why this?

Fuse.js is great in-browser for many use cases. But once data grows (for example, 10K+ records), multiple users search the same index, or mobile performance becomes a constraint, teams often move to dedicated search infrastructure.

I'm exploring whether there's room for something in between: Fuse.js ergonomics, but hosted.

Who this might be for

• Docs search
• Product catalogs
• Internal tools
• Content libraries

Who this is probably not for

• Very large-scale, enterprise-grade search systems
• Teams that need advanced relevance pipelines from day one

Early pricing direction

• Free tier for small usage
• Paid plans for higher volume

I'd love your input

1. Would this be useful to you?
2. What's your use case? (docs, catalog, internal tool, etc.)
3. What would make you pay for this vs rolling your own?
4. What's missing from this idea?

Landing page with more details: https://www.fusejs.io/cloud.html

[phinpho]

It looks great. Just a question: do you intend to make FuseCloud open source as well?

[krisk]

Fuse.js itself stays open source (Apache 2.0), always. That part isn't changing.

Fuse Cloud is where it differs. The hosted backend (indexing, the API, the infra behind it) is the commercial side. It pays for the servers and funds the time I put into the library, so no, the backend won't be open. There'll be a free tier for small usage though, so not being open source doesn't mean you have to pay to use it.

The client SDK (@fusejs/cloud) will be open though. It's just an API wrapper, and you should be able to read exactly what it does with your data.