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

# Import From iPhone

> Pull the right photos and videos off a plugged-in iPhone into Kino

Your iPhone has already analyzed every photo and video on it: faces, scenes, favorites, OCR'd text, even aesthetic scores. The Kino desktop app ships an agent skill that reads that on-device analysis directly over USB, picks the media you asked for, and imports it into Kino. It skips iCloud entirely, and you never have to hunt through thousands of files by hand.

## Run it

Plug in your iPhone, unlock it, and ask your agent:

> Pull my favorite videos from the past week off my iPhone into Kino

or filter by content:

> Find the shots of Sarah at the beach on my iPhone and bring the best ones into Kino

## What the agent does

**1. Pairs with the device**

The agent talks to the phone over USB using Apple's device protocol (it installs the open-source tooling via Homebrew if needed). If the phone hasn't trusted this Mac before, you'll get a "Trust This Computer?" prompt on the phone. Tap Trust.

**2. Reads the photo index**

Instead of downloading everything, the agent queries the phone's own Photos database: dates, favorites, people, scenes, dimensions, and quality scores. It picks candidates from metadata before transferring a single file.

**3. Pulls the selected media**

Only the files you actually want come off the phone.

**4. Imports to Kino**

The files are imported with `addMediaToKino` and AI analysis starts immediately, so they're searchable alongside the rest of your library.

## Tips

* The unlock-and-trust prompt is the most common blocker. If the agent reports it can't see the device, unlock the phone and check for the prompt.
* A nudged cable can drop the connection mid-transfer. The agent re-pairs and resumes; you don't need to start over.
* You can combine filters in one request: dates, favorites, people, scene content, and media type.


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