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

# Feedback FX

> Recursive video feedback — vertigo zoom, echo trails, or nervous jitter, in one ▶ Run.

<Frame caption="Feedback FX · screenshot coming soon">
  <img src="https://mintcdn.com/comfytv/dMkMxJDZdIo4YzAU/images/nodes/_placeholder.svg?fit=max&auto=format&n=dMkMxJDZdIo4YzAU&q=85&s=b54171099e5b02d4313fe581171ebbcc" alt="Feedback FX" width="800" height="450" data-path="images/nodes/_placeholder.svg" />
</Frame>

## What this node does

**Feedback FX** blends each frame with a transformed version of the previous output, the way a camera pointed at its own monitor creates infinite tunnels. Depending on **mode** you get a spiraling vertigo zoom, layered echo trails, or a jittery "nervous" stutter. It writes a new `COMFYTV_VIDEO` snapshot on **▶ Run**.

In and out are both `COMFYTV_VIDEO`. To feed native ComfyUI nodes, insert a **Bridge** — see the [bridge guide](https://github.com/jtydhr88/ComfyTV/blob/main/docs/bridges.md).

## When to use it

* Create psychedelic tunnels and infinite-zoom vertigo visuals.
* Leave motion echo/ghost trails behind moving subjects.
* Add a nervous, glitchy stutter to a clip for tension.

## Parameters

### mode

The feedback behavior. Options: `vertigo` (zoom/rotate spiral), `echo` (layered motion trails), `nervous` (jittery stutter). Default `vertigo`.

### phase\_increment

Rotation/phase advance per frame, `0.0`–`1.0` (step `0.005`), default `0.08`. Higher spins or shifts the feedback faster.

### zoom

Per-frame zoom of the fed-back layer, `-0.5`–`0.5` (step `0.005`), default `0.06`. Positive zooms in (tunnel inward), negative zooms out.

### feedback\_mix

How much of the previous output survives into the next frame, `0.0`–`0.98` (step `0.01`), default `0.75`. Higher = longer-lasting, deeper trails; near `0.98` the feedback almost never fades.

### style

Sub-style for the **nervous** mode. Options: `shuffle` or `scratch`. Default `shuffle`. (Only affects `nervous`.)

### frames

Frame window used to build the effect, `2`–`32`, default `32`. Larger uses more history.

### seed

Random seed, `0`–`99999`, default `7`. Same seed = repeatable jitter (mainly for `nervous`).

## Outputs

| Output    | Type            | Meaning                                               |
| --------- | --------------- | ----------------------------------------------------- |
| **video** | `COMFYTV_VIDEO` | The feedback-processed clip as a new project snapshot |

## Tips

* Keep **feedback\_mix** high for long, painterly trails; lower it if the image saturates or smears into mush.
* **style** and **seed** mainly matter in `nervous` mode.
* Small **zoom** + steady **phase\_increment** gives a smooth infinite-tunnel; larger values get chaotic fast.

## Related nodes

* **Light Graffiti** — trails built from bright pixels rather than full-frame feedback.
* **Slit Scan** — time-smear across rows/columns.