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

# Using LTX 2.3

> Install and use the LTX 2.3 audio-video plugin with ID-LoRA and IC-LoRAs in Scope

# Generating synchronized audio and video with LTX 2.3 in Scope

LTX 2.3 is a 22-billion-parameter audio-video foundation model from Lightricks. It generates synchronized video and audio from a single text prompt, which makes it different from every other pipeline shipped with Scope. This guide walks you through installing the [scope-ltx-2](https://github.com/daydreamlive/scope-ltx-2) plugin, running it locally or in the cloud, and using the two LoRA families that define this pipeline: **ID-LoRA** for identity-driven talking-head video, and **IC-LoRAs** for structural control.

<Note>
  **Running locally?** You need to install the [scope-ltx-2](https://github.com/daydreamlive/scope-ltx-2) plugin yourself (see [Install the plugin](#install-the-plugin) below). It is not bundled with Scope by default, but it can be installed directly from the desktop app.

  **Running in the cloud via [Remote Inference](/scope/guides/remote-inference)?** The plugin is already pre-installed in the cloud image. Skip the install step and jump to [Run it in the cloud](#run-it-in-the-cloud).
</Note>

<Card title="scope-ltx-2 on GitHub" icon="github" href="https://github.com/daydreamlive/scope-ltx-2">
  Source repository for the LTX 2.3 Scope plugin
</Card>

<Card title="LTX 2.3 Pipeline Reference" icon="layer-group" href="/scope/reference/pipelines/ltx-2-3">
  Full parameter schema, schedules, memory architecture, audio modes, and the complete IC-LoRA catalog. Helpful alongside this guide whenever you want the technical detail behind a setting.
</Card>

***

## Prerequisites

These requirements apply to **local** installs. For [Remote Inference](/scope/guides/remote-inference), you only need a Daydream account signed in from the desktop app, and none of the hardware or token setup below applies.

* NVIDIA GPU with 24 GB VRAM or more (RTX 4090, A5000, or similar)
* CUDA 12.8 or higher
* Python 3.12 or higher
* A HuggingFace access token with **read** permissions (see [HuggingFace Auth](/scope/guides/huggingface))

***

## Install the plugin

<Tip>
  This section applies to local installs only. If you are using [Remote Inference](/scope/guides/remote-inference), skip ahead to [Run it in the cloud](#run-it-in-the-cloud).
</Tip>

You can install the plugin from the Scope desktop app or from the CLI. Both paths install the same `ltx2` pipeline.

### Desktop app

<Steps>
  <Step title="Open the Plugins tab">
    Open **Settings** → **Plugins**.
  </Step>

  <Step title="Paste the plugin URL">
    Paste the following into the installation input field:

    ```
    https://github.com/daydreamlive/scope-ltx-2
    ```
  </Step>

  <Step title="Install">
    Click **Install** and wait for the Scope server to restart. The plugin and its `ltx2` pipeline appear automatically once the restart completes.
  </Step>
</Steps>

See the [Plugins guide](/scope/guides/plugins) for more detail on the desktop plugin workflow.

### CLI

From the scope directory:

```bash theme={null}
uv run daydream-scope install https://github.com/daydreamlive/scope-ltx-2
```

Confirm the plugin registered:

```bash theme={null}
uv run daydream-scope plugins
```

Confirm the `ltx2` pipeline is available:

```bash theme={null}
uv run daydream-scope pipelines
```

### Upgrade

In the desktop app, open **Settings** → **Plugins**. If an update is available, an **Update** button appears next to the plugin — click it and wait for the server to restart.

From the CLI:

```bash theme={null}
uv run daydream-scope install --upgrade https://github.com/daydreamlive/scope-ltx-2
```

### Uninstall

In the desktop app, open **Settings** → **Plugins** and click the trash icon next to the plugin.

From the CLI:

```bash theme={null}
uv run daydream-scope uninstall scope-ltx-2
```

***

## Run it locally

<Steps>
  <Step title="Set your HuggingFace token">
    LTX 2.3 downloads weights from the [daydreamlive/LTX2.3](https://huggingface.co/daydreamlive/LTX2.3) repository on first load. Export a token with read access before launching:

    ```bash theme={null}
    export HF_TOKEN=hf_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
    ```

    Add this to your shell profile (`~/.bashrc`, `~/.zshrc`) so it persists across sessions.
  </Step>

  <Step title="Start Scope">
    From the scope directory:

    ```bash theme={null}
    uv run daydream-scope
    ```

    The web UI opens at `http://localhost:8000`.
  </Step>

  <Step title="Select the LTX 2.3 pipeline">
    In the pipeline picker, choose **LTX 2.3**. On first load, Scope downloads roughly 28 GB of weights (transformer, text encoder, VAEs, ID-LoRA), which can take several minutes depending on your connection.
  </Step>

  <Step title="Generate">
    Enter a prompt and press **Play**. The model produces synchronized audio and video in batches of 129 frames by default.
  </Step>
</Steps>

<Tip>
  To prefetch weights without starting the full UI, run `uv run download_models --pipeline ltx2`.
</Tip>

Model weights land in `~/.daydream-scope/models/LTX2.3/` by default. Override this with the `DAYDREAM_SCOPE_MODELS_DIR` [environment variable](/scope/reference/environment-variables).

***

## Run it in the cloud

LTX 2.3 is pre-bundled in the Daydream cloud image used by [Remote Inference](/scope/guides/remote-inference), so there is no plugin install, no `HF_TOKEN` to export, and no model download to wait on. Everything that the local setup above covers is already done for you on the cloud H100.

<Steps>
  <Step title="Sign in from the desktop app">
    Open **Settings** (gear icon, top-right) → **Account** and sign in with Daydream.
  </Step>

  <Step title="Enable Remote Inference">
    Toggle **Remote Inference** on in the Account panel.
  </Step>

  <Step title="Select LTX 2.3 and press Play">
    Choose **LTX 2.3** from the pipeline picker and press **Play**.
  </Step>
</Steps>

<Card title="Remote Inference" icon="cloud" href="/scope/guides/remote-inference">
  Full Remote Inference setup, supported features, and known limitations
</Card>

***

## Generation modes

LTX 2.3 supports three modes. Parameter details for each are on the [reference page](/scope/reference/pipelines/ltx-2-3).

### Text mode

The default. A prompt produces synchronized video and audio with no other input.

### Image-to-video

Upload a reference image as the I2V input and the first frame of the generation will be conditioned on it. Use the `i2v_strength` slider to dial back the conditioning if the first frame feels too locked to the reference.

### Video mode

Used with IC-LoRAs for structural control. Select a matching IC-LoRA, switch the input to Video mode, and provide a reference video. Use `control_strength` to blend how strictly the output follows the reference.

***

## Identity-driven talking-head with ID-LoRA

ID-LoRA generates video of a specific person speaking with their own voice. Unlike cascaded pipelines that generate video and voice separately, LTX 2.3's ID-LoRA produces lip-synced audio-video in one pass.

You give it three things:

* A text prompt describing the scene
* A reference image of the subject
* A short audio clip (about 5 seconds) of the subject speaking

The output is a video of that subject speaking, in the scene you described.

### Prompt structure

ID-LoRA responds best to channel-tagged prompts:

```
[VISUAL]: A close-up of a person speaking in a park.
[SPEECH]: Hello world.
[SOUNDS]: Birds chirping.
```

Each tag controls a separate channel: `[VISUAL]` for scene content, `[SPEECH]` for what the subject says, `[SOUNDS]` for ambient audio.

### Workflow

<Steps>
  <Step title="Enable ID-LoRA mode">
    Set **Audio Mode** to `id_lora`. This tells the pipeline that the audio input is an identity reference, not a driving track.
  </Step>

  <Step title="Upload a reference audio clip">
    Upload about 5 seconds of clean audio of the subject speaking. Clip quality matters more than length.
  </Step>

  <Step title="Upload a reference image">
    Set the I2V reference image to a headshot of the same subject.
  </Step>

  <Step title="Write your prompt">
    Use the channel-tagged format above. Describe the scene, the dialogue, and the ambient audio separately.
  </Step>

  <Step title="Tune identity guidance">
    The **Identity Guidance** slider (default 3.0, range 0.0 to 20.0) controls how strongly speaker identity is amplified. Around 4.0 gives roughly 9% better speaker similarity; higher values lock identity harder at the cost of some diversity. Set it to 0 to disable.
  </Step>

  <Step title="Generate">
    Press **Play**. The output will be a video of your subject speaking the `[SPEECH]` line in the scene described by `[VISUAL]`, with `[SOUNDS]` as the environmental audio.
  </Step>
</Steps>

<Note>
  ID-LoRA weights (`ltx-2.3-id-lora-talkvid-3k.safetensors`) download automatically with the base model. You do not need to install them separately.
</Note>

***

## Structural control with IC-LoRAs

IC-LoRAs (In-Context LoRAs) condition generation on a reference video rather than a text prompt alone. They are how you get precise control over output structure: depth maps, pose skeletons, edge compositions, camera movements, or stylistic transformations like anime-to-photorealism.

### Install an IC-LoRA

<Steps>
  <Step title="Download the .safetensors file">
    Browse the [IC-LoRA catalog](/scope/reference/pipelines/ltx-2-3#ic-lora-catalog) on the reference page and download the file from HuggingFace.
  </Step>

  <Step title="Place it in the LoRA directory">
    Copy the file into `~/.daydream-scope/models/lora/`.

    ```bash theme={null}
    cp ~/Downloads/ltx-2.3-22b-ic-lora-union-control-ref0.5.safetensors ~/.daydream-scope/models/lora/
    ```

    Scope discovers new files automatically and exposes them in the LoRA picker.
  </Step>

  <Step title="Select it in the LoRA picker">
    In the Scope UI, open the LoRA adapters panel and select your IC-LoRA from the dropdown.
  </Step>

  <Step title="Switch to video mode">
    Set the input mode to Video and upload your reference video (depth map, pose skeleton, color-graded clip, or whatever the chosen IC-LoRA expects).
  </Step>

  <Step title="Tune control strength">
    Adjust `control_strength` (0.0 to 1.0) to blend the guide. Some IC-LoRAs need strengths as high as 1.3 to show the effect.
  </Step>
</Steps>

<Tip>
  IC-LoRAs expect reference video at a specific downscale factor relative to the output resolution (often 0.5x). The plugin reads this automatically from the safetensors metadata, so you do not need to configure it by hand.
</Tip>

### Which IC-LoRA should I use?

Match the IC-LoRA to your goal:

* **Depth, edges, or pose control** → Union Control (Lightricks)
* **Motion paths from spline trajectories** → Motion Track Control (Lightricks)
* **Anime-to-photorealism** → Anime2Real (community)
* **Replace regions in a video via mask** → Inpaint (community)
* **Strip color grading** → Ungrade (community)
* **Sharpen soft footage** → Refocus (community)
* **Remove compression artifacts** → Uncompress (community)
* **Extend canvas beyond the frame** → Outpaint (community)
* **Transfer camera motion** → Cameraman (community)
* **Colorize black-and-white footage** → Colorizer (community)

Full descriptions, training details, and usage notes for each are in the [IC-LoRA catalog](/scope/reference/pipelines/ltx-2-3#ic-lora-catalog).

***

## Fitting the model on a 24 GB GPU

LTX 2.3 is a 22-billion-parameter model. It runs on a 24 GB GPU through FP8 quantization of the transformer, temporary offloading of the text encoder, and CPU-resident transformer blocks that stream to GPU during denoising. This all happens automatically, but there are knobs to turn if you hit memory limits:

* Lower `num_frames` (33 or 65 instead of the default 129)
* Reduce resolution (keep to multiples of 32)
* Lower `ffn_chunk_size` to 2048 or 1024
* Set `PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True` to reduce fragmentation

See [Memory architecture](/scope/reference/pipelines/ltx-2-3#memory-architecture) for the full picture.

***

## Troubleshooting

<AccordionGroup>
  <Accordion title="Out of memory errors">
    Reduce `num_frames` first (33 or 65 instead of 129), then lower resolution, then lower `ffn_chunk_size`. Close any other GPU workloads. If the model still does not fit at any practical settings, your GPU may be below spec for LTX 2.3.
  </Accordion>

  <Accordion title="Model download fails">
    * Verify your `HF_TOKEN` is set correctly and has read access.
    * Run `huggingface-cli login` to confirm the token works.
    * Check network connectivity and available disk space (about 28 GB required).
  </Accordion>

  <Accordion title="Plugin does not appear after install">
    Restart the Scope server. From the CLI, confirm the plugin is registered with `uv run daydream-scope plugins` and that the pipeline is available with `uv run daydream-scope pipelines`.
  </Accordion>

  <Accordion title="Slow generation">
    Generation time scales with frame count, resolution, and PCIe throughput during weight streaming. Reduce `num_frames` for faster iteration and make sure no other workloads are hitting the same GPU.
  </Accordion>

  <Accordion title="ID-LoRA ignores my reference voice">
    Make sure `audio_mode` is set to `id_lora`, not `driving`. In `driving` mode, the input audio is treated as a lip-sync target and no identity transfer happens. If voice similarity is still low, raise `identity_guidance_scale` toward 4 to 6.
  </Accordion>
</AccordionGroup>

***

## See also

<CardGroup cols={2}>
  <Card title="LTX 2.3 Reference" icon="layer-group" href="/scope/reference/pipelines/ltx-2-3">
    Full parameter schema, IC-LoRA catalog, and memory architecture details
  </Card>

  <Card title="Using LoRAs" icon="wand-magic-sparkles" href="/scope/guides/loras">
    General LoRA installation and management in Scope
  </Card>

  <Card title="Using Nodes" icon="plug" href="/scope/guides/plugins">
    Install and manage third-party Scope plugins
  </Card>

  <Card title="Remote Inference" icon="cloud" href="/scope/guides/remote-inference">
    Run Scope on cloud-hosted GPUs, including LTX 2.3
  </Card>
</CardGroup>