# MemeFX

The current MemeFX MCP and Apify Actor are documented in [MEMEFX.md](MEMEFX.md). They use the 44 actual reference cuts. The following documents the separate legacy CC0 effects library.

# StrongerFX

**Sound effects for your next cut — and your AI workflow.**

254 sound effects. Six categories. Searchable metadata, waveform peaks, previews, WAV/MP3 files and explicit licence provenance.

| Category | Sounds |
|---|---:|
| Impacts | 105 |
| Interface | 72 |
| Footsteps | 25 |
| Transitions | 24 |
| Glitches | 17 |
| Notifications | 11 |

## Use with an MCP client

Requires Node.js 22 or newer. Clone this repository, then install its dependencies:

```sh
git clone https://github.com/smartpage/strongerfx-mcp.git
cd strongerfx-mcp
npm ci --ignore-scripts
```

Add a local stdio MCP server to your compatible client's configuration (replace the absolute path):

```json
{
  "mcpServers": {
    "strongerfx": {
      "command": "node",
      "args": ["/absolute/path/strongerfx-mcp/src/mcp.js"]
    }
  }
}
```

Try: **“Find three short whoosh sounds for a product reveal. Give me WAV URLs and their licences.”**

Tools:

- `search_sound_effects`: English keyword search, category, maximum duration, exact IDs, limit and format.
- `get_sound_effect`: retrieve one sound by ID.
- `list_sound_categories`: category counts.
- Resource `strongerfx://catalog`: complete metadata catalogue.

The local MCP is free, needs no API key, does not charge anything, does not execute generated code and does not write to your files. Your client can download a returned audio URL through its own tools. MCP support depends on your client; this is not a claim that every LLM supports tools.

## Apify Actor

Import this Git repository in Apify Console. Its `.actor/actor.json` declares the Dockerfile and input schema. Build, then run:

```json
{"query":"whoosh","category":"transitions","limit":3,"format":"wav"}
```

Each dataset row contains `id`, `name`, `category`, `tags`, `duration`, `sampleRate`, `channels`, `bitDepth`, `formats` (byte sizes and SHA-256), `source`, `provenance`, `downloadUrl`, `previewUrl`, `selectedFormat`, `mimeType`, and `audioBase64`.

Decode `audioBase64` to obtain the file without another download. The URL is an alternative delivery method. URLs are public; do not treat them as secrets or access controls. Results do not claim usage or viral-performance evidence.

### Pricing

The actor implements one `sound-delivered` event per output sound using the Apify SDK's budget-aware `pushData`. No matching sounds means no sound events. Maximum 50 sounds per run. The actual price and any platform usage charges are displayed in Apify Console/Store before running; source code alone does not activate monetization.

**You pay for hosted search, metadata and delivery, not exclusive rights to CC0 sounds.** These audio files are freely available in this repository; Kenney's source collections are also free. You can self-host the MCP without paying StrongerFX. No account tokens are bundled or requested in actor input.

## Licensing

- Code: MIT (see `LICENSE`).
- Kenney recordings: CC0-1.0. [Interface Sounds](https://kenney.nl/assets/interface-sounds) (100) and [Impact Sounds](https://kenney.nl/assets/impact-sounds) (130). Original licence notices are retained in `licenses/`.
- Original procedural transitions: 24 sounds generated by `scripts/build_catalog.py`, released under CC0-1.0 by StrongerFX. No third-party samples used.
- CC0 audio can be used, modified and redistributed commercially. No attribution is required, though credit to Kenney is appreciated. We do not claim exclusive ownership of Kenney's work.
- The cover illustration is promotional artwork, not part of the CC0 audio grant.

This is an editing-effects catalogue, not a collection ripped from Instagram, movies, games or famous meme recordings. “48 kHz” describes the delivered encoding; converting source recordings does not add detail absent from the originals. All WAV files are mono PCM 16-bit; MP3 versions are 192 kbps.

## Local development

```sh
npm test
npm run mcp
# Local actor smoke test (Apify SDK uses local storage):
npm start
```

Rebuilding audio requires Python 3, ffmpeg and the two source packs extracted to `assets/source-interface/` and `assets/source-impact/`. Then run `npm run catalog:build`. Original synthesis is seeded and reproducible; audio hashes can differ across ffmpeg versions.

Contact: joao@intuitiva.pt