Black Forest Labs: FLUX Video Upscale

FLUX Video Upscale enlarges an existing clip through source-faithful or creative detail enhancement. Its documented profile offers 1.5×–3× scaling, preserves the source audio and framing, and limits source length, input resolution and output frame area.

Ailoot Ultra · Video POST /v1/videos black-forest-labs/flux-video-upscale

About the model

This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.

FLUX Video Upscale is Black Forest Labs’ FLUX 3-powered video super-resolution workflow. It receives an existing clip and returns an enlarged MP4, with the source framing and audio carried into the result. A short optional description can steer the enhanced visual detail. Its two modes make a meaningful choice between staying close to the supplied appearance and generating more aggressive detail, rather than treating every larger image as an equally faithful reconstruction.

Choose precise or creative enhancement

The creativity setting selects the mode. A value of 0 is the precise, source-faithful profile: the manufacturer recommends it when a face, product or brand asset needs to keep its identity. A value of 1 enables creative detail enhancement and is the default. It can restore or invent texture more aggressively, which is useful for generated footage, scenery, crowds or surfaces whose detail can be developed creatively.

The maker explicitly warns that creative enhancement does not preserve faces and products as strictly. Check the feature that must remain identifiable: facial appearance, a product outline, lettering, a material pattern or a logo. A visual that looks sharper can still contain a changed or newly invented detail. For a product demonstration, use the intended product appearance as the review criterion rather than selecting a version only because its texture is more elaborate.

The optional prompt describes the content of the clip to guide enhancement; an empty description requests a neutral upscale. An original generated-scene brief could identify a moss-covered stone wall, soft daylight and fine leaf texture. Keep the wording about the material that should be refined. A request to replace an object, rebuild the setting or change dialogue belongs to an editing decision, while this profile retains the source audio track.

Scale and output frame area

The documented upscale_factor ranges from 1.5 to 3, with 2 as the default. It enlarges the source resolution while preserving the aspect ratio. The selected factor and the input dimensions determine the intended output, so a landscape clip and a portrait composition need to be planned from their actual frames. Prepare the target display size before choosing how much enlargement is useful.

The exact specification caps each output frame at 13.75 of the manufacturer’s megapixel units. Its documentation defines one such unit as 1,048,576 pixels, giving a ceiling of 14,417,920 pixels, approximately 14.4 million. That explains the rounded 14.4-megapixel wording in the guide. A large source can therefore receive less enlargement than the requested factor. Check the delivered dimensions rather than assuming that a 3× request always produces three times each source dimension.

The published source-resolution ceiling is 2560×1440. This workflow increases an eligible source toward 4K and beyond; it does not accept an already larger source simply because a higher output is desired. Downscale or prepare an appropriate working copy before using that documented profile. Keep the original recording separately so the final comparison is not limited to a resized intermediate.

Prepare a short, readable source

The maker’s ordinary preparation guidance limits sources to twenty seconds and fifty MB. A clip is supplied through a video URL or encoded MP4. Its exact reference notes that processing covers the first twenty seconds: a small length overshoot can be processed only to that mark, while a substantially longer clip is rejected. No numerical tolerance is published. Trim to the intended twenty-second-or-shorter segment instead of depending on that unspecified boundary behavior.

Choose the least compressed version available. Compression artifacts limit the amount of real detail the upscaler can recover, and creative enhancement can make a plausible texture without establishing what the original scene contained. Complete the intended edit and trim before enlarging the selected clip. An original workflow might first finish a product shot, choose the useful action interval and then upscale that final selection, retaining the unmodified source for comparison.

The result keeps the source audio track. Inspect its timing with the picture, especially at the beginning and end of a selected segment. The documented profile does not supply an independent frame-rate selector, new dialogue generation or image-reference editing controls. Its purpose is the larger video and chosen treatment of detail, with the source’s temporal and sonic material retained within the processed segment.

Keep and check the delivered clip

The producer workflow is asynchronous: a submitted task supplies an identity and status location, and its completed result contains a signed MP4 delivery link. That link expires after about one hour. Retain the actual clip, source, selected factor, creativity mode and content description instead of relying on the temporary location. Keep failed or moderated status distinct from a completed video.

Review individual frames for faces, markings and surface changes, then watch the sequence for continuity, motion and audio alignment. Precise and creative outputs answer different visual requirements; the useful artifact is the enlarged clip that meets the chosen fidelity goal at the intended display size. A larger resolution alone is not evidence that the original identity, wording or fine detail remained correct.

Ailoot Ultra · Video POST /v1/videos black-forest-labs/flux-video-upscale

black-forest-labs/flux-video-upscale

Black Forest Labs /black-forest-labs/flux-video-upscale

Overview

Accepted inputs: Text · Images · Audio · Video
Results: Video

Available in Ailoot · inputs & methods

Ailoot Ultra · Video POST /v1/videos

Input types: text, image, audio, video
Output types: video

Field	Type	Requirement	Values & limits	Default
Field`model`	Type`string`	RequirementRequired	Values & limits black-forest-labs/flux-video-upscale	Default—
Field`input_references`	Type`array`	RequirementRequired	Values & limits Minimum items: 1 Maximum items: 1	Default—
Field`upscale_factor`	Type`number`	RequirementRequired	Values & limits Minimum: 1.5 Maximum: 3	Default—
Field`creativity`	Type`integer`	RequirementRequired	Values & limits 0, 1	Default—

Method parameters and lifecycle

Parameter	Contract
Parameter`content_type`	Contractapplication/json
Parameter`required`	Contractmodel, input_references, upscale_factor, creativity
Parameter`material_types`	Contracttext, image, audio, video
Parameter`material_conditions`	ContractUse exactly one actual completed owned video asset content URL with its returned signature. The asset owner, type and stored byte witness must match the request.

Lifecycle: asynchronous_job_then_owned_outputs

GET /v1/videos/{request_id} · actual returned id or polling_url

GET /v1/videos/{request_id}/content

Pricing

How the final charge is calculated

The displayed rates estimate the cost before generation. The final charge is calculated from the cost recorded for the original generation. Measured output duration and forecast quantities are not the final billable quantity.

Tariff references do not add request options. Supported parameters are listed for the selected method.

Fast mode: Not specified When fast mode is selected
Flex mode: Not specified When flex mode is selected
Maximum reasoning effort: Not specified When maximum reasoning effort uses the default adjustment

Estimated rates for comparison

Prices by configuration

Match each charge to the conditions shown. Use one rate for each set of alternative configurations.

Configuration	Rate schedule	Forecast quantity	Billing unit	Billing conditions	Pixels per megapixel	Charge
ConfigurationBase token rates	Rate scheduleBase rate	Forecast quantity—	Billing unit—	Billing conditions—	Pixels per megapixel—	Charge Input · Input tokens · Estimate for comparison only — / 1K tokens Output · Output tokens · Estimate for comparison only — / 1K tokens Input · Cache read · Estimate for comparison only — / 1K tokens Input · Cache write · Estimate for comparison only — / 1K tokens Input · Cache write, 1 hour · Estimate for comparison only — / 1K tokens
ConfigurationVideo price	Rate schedule—	Forecast quantity—	Billing unitper megapixel-second	Billing conditionsquality: creative	Pixels per megapixel1,048,576	Charge Output · Video price · Estimate for comparison only $0.21 per megapixel-second
ConfigurationVideo price	Rate schedule—	Forecast quantity—	Billing unitper megapixel-second	Billing conditionsquality: precise	Pixels per megapixel1,048,576	Charge Output · Video price · Estimate for comparison only $0.15 per megapixel-second
ConfigurationEstimated generation cost	Rate schedule—	Forecast quantityone generation estimate	Billing unit—	Billing conditions—	Pixels per megapixel—	Charge Output · Estimated generation cost · Estimate for comparison only $0.15 estimate before generation This estimate is not an additional charge or a fixed rate for the completed generation.

Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →

Example requests

Ailoot Ultra · Video

Use actual input and retain the complete result

set -eu
: "${AILOOT_API_KEY:?Set your Ailoot API key}"
: "${AILOOT_INPUT_REFERENCES_JSON:?Set the actual documented input value}"
export AILOOT_INPUT_REFERENCES_JSON
: "${AILOOT_UPSCALE_FACTOR:?Set the actual documented input value}"
export AILOOT_UPSCALE_FACTOR
: "${AILOOT_CREATIVITY:?Set the actual documented input value}"
export AILOOT_CREATIVITY
python3 - <<'PY_REQUEST'
import json,os
from pathlib import Path
request={'model': 'black-forest-labs/flux-video-upscale', 'input_references': '${ACTUAL_OWNED_INPUT_REFERENCES}', 'upscale_factor': '${ACTUAL_UPSCALE_FACTOR}', 'creativity': '${ACTUAL_CREATIVITY}'}
request['input_references']=json.loads(os.environ['AILOOT_INPUT_REFERENCES_JSON'])
request['upscale_factor']=float(os.environ['AILOOT_UPSCALE_FACTOR'])
request['creativity']=int(os.environ['AILOOT_CREATIVITY'])
Path('request.json').write_text(json.dumps(request,ensure_ascii=False),encoding='utf-8')
PY_REQUEST
curl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/videos \
  -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \
  --data-binary @request.json -D response-headers.txt -o response.bin
# Retain the complete response and headers, every actual output, original returned identifiers and all usage fields. Recover the original operation without submitting creation again.

Receive and save results

Ailoot Ultra · Video

GET /v1/videos/{request_id}
GET /v1/videos/{request_id}/content

Save all completed assets

GET all completed assets

set -eu
: "${AILOOT_API_KEY:?Set your Ailoot API key}"
: "${AILOOT_STATUS_FILE:?Set the actual completed status JSON file}"
export AILOOT_STATUS_FILE
python3 - <<'PY_REQUEST'
import json,os
from pathlib import Path
from urllib.parse import urlsplit,urljoin
f=Path(os.environ["AILOOT_STATUS_FILE"])
if not f.is_file():raise SystemExit("Choose the actual completed status file")
r=json.loads(f.read_text())
if r.get("status")!="completed" or not isinstance(r.get("data"),list) or not r["data"]:raise SystemExit("Wait for a completed status with asset paths")
urls=[]
for x in r["data"]:
    if not isinstance(x,str) or "\n" in x or "\r" in x:raise SystemExit("Invalid asset path")
    u=urljoin("https://api.ailoot.ai",x)
    z=urlsplit(u)
    if z.scheme!="https" or z.netloc!="api.ailoot.ai" or z.username or z.password:raise SystemExit("Use the actual Ailoot asset paths from status")
    urls.append(u)
Path("asset-urls.txt").write_text("\n".join(urls)+"\n")
PY_REQUEST
index=0
while IFS= read -r asset_url; do
  curl --fail-with-body --silent --show-error "$asset_url" \
    -H "Authorization: Bearer $AILOOT_API_KEY" -o "video-${index}.mp4"
  index=$((index + 1))
done < asset-urls.txt

Keep all completed results

Save every completed video or returned file; retrieve every actual returned asset URL.

Retain the complete response and headers, every actual output, original returned identifiers and all usage fields. Recover the original operation without submitting creation again.

If the original response needs checking

Keep X-Client-Request-ID, X-Ailoot-Task-Id and actual returned id/polling_url. Continue the original job; do not submit creation again.

Response fields

Field	Type
Field`id`	Type
Field`status`	Type
Field`data[]`	Type
Field`outputs[]`	Type

Aliases

Model aliases

Ailoot Ultra black-forest-labs/flux-video-upscale

Method path aliases

Use the method paths shown above.

Related caller IDs

black-forest-labs/flux-video-upscale

FAQ · Black Forest Labs: FLUX Video Upscale

What inputs and results are available?: POST /v1/videos · black-forest-labs/flux-video-upscale: text, image, audio, video → video
Which fields does the request use?: POST /v1/videos · black-forest-labs/flux-video-upscale: model, input_references, upscale_factor, creativity
What limits apply?: POST /v1/videos · black-forest-labs/flux-video-upscale: upscale_factor.minimum: 1.5; upscale_factor.maximum: 3

Black Forest Labs: FLUX Video Upscale