.. _genai:
AI Generator
============
smartElements offers an AI generator that lets you generate AI videos and
images. In order to generate genAI output you will need a running internet
connection and a `replicate `_ token for
authentication.
The following shows you a quick overview how to use genAI inside smartElements
to generate your own elements. Let's generate some fire and explosion videos
and images.
.. raw:: html
|
Launching the AI generator
--------------------------
You can launch the AI generator Inside smartElements using the AI generator
button which you can find on the top right:
.. image:: img/genai/launch_genai.png
The AI generator launches:
.. image:: img/genai/ai_generator.png
It contains the following sections:
**1) AI form:** Here you edit all details for the output to be generated.
You can switch between video and image models and set up all options
and enter a prompt. When you are ready click the blue **Run** button at
the bottom to generate a new output.
**2) Output:** Here you can view the generated output. This section can
show video and image outputs. Some models like flux schnell are able to
generate multiple outputs which you can also cycle through here as well.
**3) Settings:** Click here for updating the AI generator's settings.
**4) Search:** Here you can filter outputs by used words in the prompts.
**5) Output table:** Here you see all generated outputs. The newest outputs
are listed top, the oldest at the bottom.
Setting up an API token
-----------------------
Before you can generate new outputs, you need to set up a replicate account
and generate a token. A token is necessary to use replicate's services and so
that it can identify you. For that, click the settings button **(3)** at the
top right as also seen in the previous image. The settings section shows up:
.. image:: img/genai/settings.png
The settings contains the following sections:
**1) API token:** Here you can set the replicate API token. This works
like a password and lets replicate identify you.
**2) Output root directory:** The root directory where all GenAI outputs
are downloaded to. By default, it points to the smartElements folder in
your home directory. Change this if needed to another root directory.
Click the **Browse API tokens** button to reach replicate's website. Here you
can create a new account. Once you have set it up and entered your credit card
details, hover over the 'Default' token (or create a new token if you like) and
copy its value to your clipboard. In the settings you can then paste it.
Click the **Save settings** button at the bottom to apply the changes.
Generating new outputs
----------------------
The left section lets you enter your prompt and edit all other options the
currently selected model offers. The offered options change depending on the
currently selected model.
smartElements offers the following video and image models, which we will extend
in the future.
Video models:
- google/veo-2
- pixverse/pixverse-v4.5
Image models:
- black-forest-labs/flux-1.1-pro-ultra
- black-forest-labs/flux-1.1-pro
- black-forest-labs/flux-schnell
.. image:: img/genai/edit_form.png
:scale: 60
When you have set up your prompt and model options to your needs, click the
**Run** button. This will kick off a new output generation. Depending on the
chosen model and and its output type, this might take a few seconds.
Once done, you can view it in the output section on the right:
.. image:: img/genai/viewing_output.png
Right click any item in the output table:
.. image:: img/genai/context_menu.png
:scale: 60
The context menu contains the following options:
**1) Reveal:** Reveal the output in your explorer.
**2) Tweak it:** Set up the left section with the exact options + prompt
that were used to generate that selected output. This is useful if you
want to get back to a particular output setting and either tweak it or
create variants from it.
**3) Delete:** Click to delete the output and all its metadata.
**Please note:** There is no undo.
Ingesting outputs
-----------------
In order to use the elements in your working file, you need to ingest these to
bring them to smartElements. You can send them to ingestion directly from the
AI generator. Simply select the outputs that you want to ingest and click the
**Ingest selected** button at the bottom.
.. image:: img/genai/send_to_ingest.png
The ingestion matrix launches. Set the list(s) to ingest to and adjust other
things to your needs, like tags and the transcoding processor, etc. For
more information on the ingestion matrix please have a look at the
`ingestion `_ section.
.. image:: img/genai/ingesting.png
Once the ingestion is done, click the reveal button. This navigates you
directly to the ingested output which is then ready to be used.
.. image:: img/genai/smartElements_genai_media_overview.png