# Getting Started
URL: /docs/getting-started

Requirements and the shortest path to a working search index.



## Requirements [#requirements]

Tachyon ships as a single Linux binary inside a small Docker image (Alpine,
musl libc). To run it you need:

* Docker, **or**
* a Rust toolchain if you'd rather build from source ([Installation](/docs/installation))

There's no separate database, no JVM, and no external dependencies to
provision.

## The workflow [#the-workflow]

Every Tachyon deployment follows the same four steps:

1. **Run Tachyon** — start the binary, pointed at a data directory
2. **Create a collection** — declare the fields you want to search, filter, sort, or facet on
3. **Index documents** — send flat JSON objects to be indexed
4. **Search** — query over HTTP and get ranked results back

```bash
docker run -p 8108:8108 -v tachyon-data:/data adikeshri/tachyon
```

This starts Tachyon with no API key configured, which is fine for local
development and not fine on a network anyone else can reach — see
[Configuration](/docs/configuration#authentication) before deploying it
anywhere shared.

## First search [#first-search]

Continue to the [Quickstart](/docs/quickstart) for a complete, copy-pasteable
walkthrough: creating a `products` collection, indexing two documents, and
running your first search — including a misspelled one.