> ## Documentation Index
> Fetch the complete documentation index at: https://docs.eversince.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Cloud processing and storage

> What work on Eversince's servers uses in minutes, what each plan includes, and what more costs.

Cloud processing is paid in minutes. A minute is a unit of work on Eversince's servers, not a minute of clock time. Through the desktop app, reading files, transcribing, detecting beats, speakers and loudness, indexing a video, reframing, syncing recordings, drawing frames, and rendering run on the user's computer at no charge. [Desktop and cloud](/desktop-and-cloud) has the rule.

## What each plan includes

<div className="price-table">
  | Plan | Cloud processing | Storage |
  | - | - | - |
  | Free | 25 minutes, once | 5 GB |
  | Pro | 500 minutes a month | 100 GB |
  | Business | 1,000 minutes a month per member | 200 GB per member, shared by the team |
</div>

On a paid plan, $10 buys 500 more minutes, the smallest purchase, and bought minutes last 12 months. More storage is $5 a month or \$48 a year per 100 GB.

## What work uses

<div className="price-table">
  | Work | Tool | For | Minutes |
  | - | - | - | - |
  | Transcribe speech | `transcribe_media` | An hour of audio | 22 |
  | Index a video for search | `search_footage` | An hour of video | 8.2 |
  | Write a video's shot log | `inspect_media` | 10 minutes of video | 8.2 |
  | Find the cuts in a video | `inspect_media` | 10 minutes of 1080p video | 3.2 |
  | Ask a model about a video | `inspect_media` | 10 minutes of video | 11 |
  | Recognize speakers | `speakers` | An hour of audio | 11 |
  | Find where speech is, for removing silence, ducking music or cutting to the speaker | `remove_silence` | An hour of audio | 2.7 |
  | Remove background noise | `clean_audio` | An hour of audio | 7.2 |
  | Isolate the voice | `clean_audio` | 10 minutes of audio | 180 |
  | Level loudness | `level_audio` | An hour of audio | 1.4 |
  | Detect beats | `detect_beats` | A 4-minute track | 0.4 |
  | Sync cameras by sound | `sync_recordings` | Two 1-hour recordings | 0.1 |
  | Keep the subject in frame | `auto_reframe` | A minute of 1080p video | 0.1 |
  | Cut a subject out of a photo | `edit_canvas` | One 12-megapixel photo | 0.1 |
  | Mask a person on a timeline clip | `edit_timeline` | A minute of 1080p video | 5 |
  | Draw a frame of a timeline or canvas | `inspect_timeline` | One frame | 0.2 |
  | Render a canvas as a PDF | `render_canvas` | One page | 0.1 |
  | Render a timeline | `render_timeline` | A minute at 1080p | 3.3 |
  | Render a timeline | `render_timeline` | A minute at 4K | 9.7 |
  | Cut the part a clip plays into its own file, for upscaling or removing its background | `upscale_media` | A minute of 1080p video | 0.9 |
  | Describe an uploaded picture so it can be searched | `import_media` | One picture | 0.2 |
  | Convert a video to play everywhere, on a paid plan | `convert_media` | A minute of 1080p video | 4.2 |
</div>

For example, transcribing a 30-minute interview uses 11 minutes.

A figure is the estimate `estimate_cost` gives before the work. Transcription and voice isolation are charged as shown; other work is charged what it actually cost, more or less than the estimate. A render with effects, captions or transitions costs more than a plain cut, and a render that uses up the minutes left stops and is charged for what ran. Transcripts, indexes and shot logs are kept, so reading them again is free.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.