# TTS Workshop

Search the Tabletop Simulator Steam Workshop, fetch full TTS save files, and
analyze their contents. A lightweight, client-only pnpm monorepo.

## Packages

| Package             | Role                                                   | Runtime    |
| ------------------- | ------------------------------------------------------ | ---------- |
| `apps/proxy`        | Hono HTTP server: Workshop search + save fetch         | Node       |
| `packages/tts`      | Fetch save from Steam, BSON-parse to `TTSMod`          | Node       |
| `packages/extract`  | Analyze a `TTSMod`: objects, asset refs, downloads     | Isomorphic |
| `packages/shared`   | Shared types + zod schemas                             | Isomorphic |

See [`docs/architecture.md`](docs/architecture.md) for the architecture and
[`docs/implementation-plan.md`](docs/implementation-plan.md) for the plan.

## Setup

```sh
pnpm install
cp .env.example .env   # then set STEAM_API_KEY
pnpm dev               # runs the proxy at http://localhost:3000
```

Get a Steam Web API key at https://steamcommunity.com/dev/apikey (free).

## API

| Method | Path              | Description                                  |
| ------ | ----------------- | -------------------------------------------- |
| GET    | `/health`         | Liveness                                     |
| GET    | `/search?q=&page=`| Search the Workshop (scrapes browse page)    |
| GET    | `/items/:id`      | Full parsed `TTSMod` (BSON save)             |
| GET    | `/items/:id/file` | Raw save bytes, filename from header         |

## Commands

```sh
pnpm dev          # run the proxy (tsx watch)
pnpm build        # compile all packages
pnpm typecheck    # typecheck all packages
pnpm lint         # lint all packages
```

## How search works

Steam has no official search API. The proxy fetches the Workshop browse page
(`steamcommunity.com/workshop/browse/?appid=286160`) and parses the embedded
`window.SSR.renderContext` JSON (a React Query cache containing a
`workshop_browse` entry with the results). This is more robust than scraping
the DOM, but Steam can still change the page structure — if search breaks, that
parser is the first place to look.

## Notes

- `STEAM_API_KEY` is only needed for `/items/*` (resolving `file_url`). Search
  works without it.
- Each request fetches fresh; there is no caching by design (client-only tool).
- `packages/extract` has zero runtime dependencies and runs in browser or Node,
  ready for a future frontend.