FLUX 3 First Last Frame to Video API

blackforestlabs/flux-3/first-last-frame-to-video

FLUX 3 First Last Frame to Video anchors both opening and closing still frames to generate 5–20 seconds of coherent transitional video with optional synchronized native audio. Direct motion trajectories, camera movement, and audio between two precise visual compositions.

Input

Prompt

457/20000

Start frame

Drag and drop files, paste from clipboard, or provide a URL.

1/1 files added

End frame

Drag and drop files, paste from clipboard, or provide a URL.

1/1 files added

Duration

Resolution

Aspect ratio

Sound

Output

Ready

Continue with

FLUX 3 Text to Video FLUX 3 Image to Video FLUX 3 Extend Video FLUX 3 Keyframes to Video

Examples

Begin exactly on the first frame and finish on the last frame. One continuous 5-second locked close-up: the folded paper crane's wings slowly uncrease and lift until they match the end still. Preserve the same crane, paper color, table, window light, and camera. Native audio: dry paper flex, a faint wooden-table creak, quiet room tone. No logos, no readable text, no products, no packaging, no prices, no CTA, no advertising, no watermark, no brand marks.

Begin on the first frame and finish on the last frame. One continuous 5-second studio shot: the dancer rises from the low opening stance into the tall closing pose with one smooth turn. Preserve the same dancer, practice clothes, wooden floor, and soft side light. Camera: locked full-body wide shot. Native audio: bare feet on wood, steady breath, room tone. No logos, no readable text, no products, no packaging, no prices, no CTA, no advertising, no watermark, no brand marks.

Begin on the first frame and finish on the last frame. One continuous 5-second locked wide shot of the same rooftop garden: dawn blue light warms into gold morning sun, plant leaves lift gently in a rising breeze. Preserve the same pots, railing, skyline, and camera. Native audio: morning birds, a distant city hum, soft wind. No logos, no readable text, no products, no packaging, no prices, no CTA, no advertising, no watermark, no brand marks.

REST API

Quick Start

Authenticate with the API, supply opening and closing frame URLs, and retrieve video using the task ID.

Connect to the Vidgo API

Create an API key, store securely on your server, and include Authorization: Bearer VIDGO_API_KEY.

Endpoint: POSThttps://api.vidgo.ai/api/generate/submit
Authentication: Authorization: Bearer VIDGO_API_KEY

Submit one generation task

Fill parameters with start and end URLs in order, submit, and save task_id for progress queries.

Submit request

REQUEST_BODY=$(cat <<'JSON'
{
  "model": "blackforestlabs/flux-3/first-last-frame-to-video",
  "input": {
    "prompt": "Begin exactly on the first frame and finish on the last frame. One continuous 5-second locked close-up: the folded paper crane's wings slowly uncrease and lift until they match the end still. Preserve the same crane, paper color, table, window light, and camera. Native audio: dry paper flex, a faint wooden-table creak, quiet room tone. No logos, no readable text, no products, no packaging, no prices, no CTA, no advertising, no watermark, no brand marks.",
    "duration": 5,
    "resolution": "720p",
    "aspect_ratio": "auto",
    "sound": true,
    "image_urls": [
      "https://cdn.vidgo.ai/apis/models/blackforestlabs/flux-3/first-last-frame-to-video/v1/01/input-start-frame.webp",
      "https://cdn.vidgo.ai/apis/models/blackforestlabs/flux-3/first-last-frame-to-video/v1/01/input-end-frame.webp"
    ]
  }
}
JSON
)

RESPONSE=$(curl --silent --show-error --fail-with-body \
  --request POST \
  --url "https://api.vidgo.ai/api/generate/submit" \
  --header "Authorization: Bearer $VIDGO_API_KEY" \
  --header "Content-Type: application/json" \
  --data "$REQUEST_BODY")

CODE=$(printf '%s' "$RESPONSE" | jq -r '.code // empty')
if [ "$CODE" != "0" ] && [ "$CODE" != "200" ]; then
  printf 'API error: %s
' "$RESPONSE" >&2
  exit 1
fi
printf '%s
' "$RESPONSE"

Wait for the result

Poll with task_id while running, stop on finished or failed. On success, download data.files[].file_url.

Track status

GET https://api.vidgo.ai/api/generate/status/{task_id}

Poll status with a 2-second base interval. Continue querying while not_started or running, and stop on finished or failed. You can also supply callback_url in the request to receive webhook notifications.

not_startedrunningfinishedfailed

Submit response

{
  "code": 200,
  "data": {
    "task_id": "task-flux3-flf-...",
    "status": "running",
    "created_time": "2026-09-16T10:00:00Z"
  }
}

Finished response

{
  "code": 200,
  "data": {
    "task_id": "task-unified-...",
    "status": "finished",
    "files": [
      {
        "file_url": "https://storage.vidgo.ai/generated/video.mp4",
        "file_type": "video"
      }
    ],
    "created_time": "2026-08-22T10:00:00Z",
    "progress": 100,
    "error_message": null
  }
}

Complete runnable example

Expand for an end-to-end integration script with response checks, polling loops, and error boundaries.

Complete runnable example

set -euo pipefail
: "${VIDGO_API_KEY:?Set VIDGO_API_KEY in your environment}"

REQUEST_BODY=$(cat <<'JSON'
{
  "model": "blackforestlabs/flux-3/first-last-frame-to-video",
  "input": {
    "prompt": "Begin exactly on the first frame and finish on the last frame. One continuous 5-second locked close-up: the folded paper crane's wings slowly uncrease and lift until they match the end still. Preserve the same crane, paper color, table, window light, and camera. Native audio: dry paper flex, a faint wooden-table creak, quiet room tone. No logos, no readable text, no products, no packaging, no prices, no CTA, no advertising, no watermark, no brand marks.",
    "duration": 5,
    "resolution": "720p",
    "aspect_ratio": "auto",
    "sound": true,
    "image_urls": [
      "https://cdn.vidgo.ai/apis/models/blackforestlabs/flux-3/first-last-frame-to-video/v1/01/input-start-frame.webp",
      "https://cdn.vidgo.ai/apis/models/blackforestlabs/flux-3/first-last-frame-to-video/v1/01/input-end-frame.webp"
    ]
  }
}
JSON
)

SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  --request POST \
  --url "https://api.vidgo.ai/api/generate/submit" \
  --header "Authorization: Bearer $VIDGO_API_KEY" \
  --header "Content-Type: application/json" \
  --data "$REQUEST_BODY")

TASK_ID=$(printf '%s' "$SUBMIT_RESPONSE" | jq -r '.data.task_id // .task_id // empty')
BUSINESS_CODE=$(printf '%s' "$SUBMIT_RESPONSE" | jq -r '.code // empty')
if [ "$BUSINESS_CODE" != "0" ] && [ "$BUSINESS_CODE" != "200" ]; then
  printf 'Submit failed:
%s
' "$SUBMIT_RESPONSE" >&2
  exit 1
fi
if [ -z "$TASK_ID" ]; then
  printf 'Submit response did not include task_id:
%s
' "$SUBMIT_RESPONSE" >&2
  exit 1
fi

START_TIME=$(date +%s)
POLL_DELAY=2
while true; do
  if [ $(( $(date +%s) - START_TIME )) -ge 600 ]; then
    printf 'Timed out after 600 seconds
' >&2
    exit 1
  fi
  STATUS_RESPONSE=$(curl --silent --show-error --fail-with-body \
    --url "https://api.vidgo.ai/api/generate/status/$TASK_ID" \
    --header "Authorization: Bearer $VIDGO_API_KEY")
  STATUS=$(printf '%s' "$STATUS_RESPONSE" | jq -r '.data.status // .status // empty')
  BUSINESS_CODE=$(printf '%s' "$STATUS_RESPONSE" | jq -r '.code // empty')
  if [ "$BUSINESS_CODE" != "0" ] && [ "$BUSINESS_CODE" != "200" ]; then
    printf 'Status request failed:
%s
' "$STATUS_RESPONSE" >&2
    exit 1
  fi

  case "$STATUS" in
    finished)
      printf '%s' "$STATUS_RESPONSE" | jq -r '(.data.files // .files // [])[]?.file_url'
      break
      ;;
    failed)
      printf '%s' "$STATUS_RESPONSE" | jq -r '.data.error_message // .error_message // "Generation failed"' >&2
      exit 1
      ;;
    not_started|running)
      sleep "$POLL_DELAY"
      if [ "$POLL_DELAY" -lt 10 ]; then POLL_DELAY=$((POLL_DELAY + 1)); fi
      ;;
    *)
      printf 'Unexpected task status: %s
' "$STATUS" >&2
      exit 1
      ;;
  esac
done

Input parameters

The table lists available input parameters, types, and defaults. Request examples also include the required top-level model field. Prepare inputs and configure output specifications.

Field	Type	Required	Default	Description
prompt	string	Yes	—	Describes transition motion, camera movement, and audio between frames.
image_urls	string[]	Yes	—	Array of exactly 2 public image URLs. Index 0 is start frame, index 1 is end frame.
duration	integer	No	5	Output video duration in seconds from 5 to 20.
resolution	string	No	720p	Output resolution, 720p or 1080p.
aspect_ratio	string	No	auto	Framing ratio: auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, or 9:16.
sound	boolean	No	true	Whether to generate native synchronized audio.

Response Fields

A successful submission returns a task ID. Status queries provide progress, output files, and error details when a task fails.

Field	Type	Description
code	integer	Application result code; successful responses use 0 or 200.
message	string	Human-readable message or error detail when present.
data.task_id	string	Task identifier used for querying generation status.
data.status	string	not_started, running, finished, or failed.
data.created_time	string	Task creation timestamp in date-time format.
data.progress	integer	Progress percentage reported from 0 to 100.
data.files[]	array	Array of output files generated upon task completion.
data.files[].file_url	string	Public URL for the generated video file.
data.files[].file_type	string	Output file MIME category, such as video.
data.error_message	string \| null	Failure details when status is failed.

Task Lifecycle

Continue querying while status is not_started or running. End polling at finished or failed, then process output files or error details respectively.

not_started
The task was accepted and is waiting in the queue.
running
Generation is in progress. Continue polling the task_id.
finished
Generation succeeded. Read video URLs from data.files[].file_url.
failed
Generation failed. Read data.error_message and halt polling.

Polling and Errors

AuthenticationOn 401 response, verify the Bearer API key in the Authorization header and retry.
ValidationOn 400 response, verify that image_urls has exactly 2 public URLs in chronological order.
Network and timeoutIf status polling encounters network interruptions, retain task_id and retry status queries.
Polling intervalPoll status with a 2-second base interval, extending intervals for longer renders.
Terminal statesContinue polling only on not_started or running. Stop immediately once finished or failed.
Callback optionProvide callback_url in request top level to receive final task payloads via webhook.

Endpoint limits

Specification	Value	Details
Input mode	2 images (start/end) + Prompt	Ordered pair of start and end image URLs with transition prompt.
Output	Video with native audio	Returns asynchronous task ID yielding standard MP4 video.
Duration	5–20 seconds	Integer range, default 5 seconds.
Resolution	720p / 1080p	Default is 720p.
Aspect ratio	8 options (including auto)	auto, 21:9, 2:1, 16:9, 4:3, 1:1, 3:4, 9:16, default auto.
Billing basis	Duration × Resolution rate	34 credits/sec for 720p, 58 credits/sec for 1080p.

View FLUX 3 First Last Frame to Video API documentation

FLUX 3 First Last Frame to Video

FLUX 3 First Last Frame to Video provides precise boundary control over both the start and culmination of a shot. Supplying ordered start and end frames establishes visual targets, allowing the model to compute plausible physical motion, subject transitions, and camera movement over 5–20 seconds with native synchronized audio.

Why Choose This?

Dual anchor composition controlFirmly pins down both opening pose and closing composition, guiding the video toward your intended final frame.
Smooth physical interpolationLeverages deep motion dynamics to generate organic displacement, material deformations, and fluid transitions between two still baselines.
Prompt-guided transition choreographyUse text instructions to guide intermediate details such as camera pacing, turning points, and progressive lighting changes across the clip.
Native synchronized sound synthesisGenerates audio that follows the arc of movement, synchronizing footfalls, ambient acoustic shifts, and music to visual progression.
Configurable 5–20 second pacingAllows fine-tuning generation length from 5 to 20 seconds, accommodating both brisk visual cuts and slow, contemplative morphs.
Pristine 720p and 1080p deliveryMaintains crisp sharpness and textural depth from opening frame to closing frame across 720p and 1080p resolution choices.

Parameters

Parameter	Requirement	Description
prompt	Required	String. Directs motion, camera trajectory, transitional behavior, and native audio connecting the two frames.
image_urls	Required	String array containing exactly 2 public image URLs. Item 0 is the start frame, and item 1 is the end frame.
duration	Optional	Integer. Sets output video duration in seconds from 5 to 20; the Playground preselects 5 seconds. Default`5`
resolution	Optional	String. Sets output resolution to 720p or 1080p; the Playground preselects 720p. Default`720p1080p`
aspect_ratio	Optional	String. Controls framing ratio, supporting auto and standard formats; the Playground preselects auto. Default`auto21:92:116:94:31:13:49:16`
sound	Optional	Boolean. Controls whether native synchronized audio is generated alongside video; default is true. Default`truefalse`

How to Use

Supply start and end image URLsPrepare publicly accessible URLs for the opening still and closing still, ordered in image_urls as [start_url, end_url].
Describe transition and camera behaviorIn prompt, detail the movement trajectory connecting both states, specifying camera maneuvers and audio cues.
Set duration and timingSelect an output duration between 5 and 20 seconds depending on whether the transition requires rapid or gradual pacing.
Choose resolution and aspect ratioSelect 720p or 1080p resolution and configure an appropriate aspect ratio or leave set to auto.
Configure audio synthesisLeave sound set to true to synthesize matching transitional audio, or toggle to false for silent output.
Generate and review transitionConfirm the estimated credits, click Run, and inspect the continuous motion linking the two images before downloading.

Pricing

Billed per output second by resolution tier, including native synchronized audio. 1 credit = $0.005.

Usage	Rate	Details
720p	34 credits / sec ($0.17 / sec)	Default 5s at 720p is 170 credits ($0.85).
1080p	58 credits / sec ($0.29 / sec)	5s at 1080p is 290 credits ($1.45).

Best Use Cases

Storyboard scene transitionsConnect discrete storyboards by computing natural continuous camera movements between key compositions.
Product transformation revealsIllustrate product state changes, unboxing, or mechanical assembly by setting closed and open stills as boundary anchors.
Character posture choreographyDefine start and end stances for characters, allowing the model to synthesize balanced physical locomotion.
Seamless looping animationsPass identical opening and closing frames to create continuous, cyclic dynamic loops.

Pro Tips

Keep subject appearance, wardrobe, and ambient perspective logically consistent across both stills for seamless interpolation.
Focus prompt instructions on how the subject transitions, such as 'subject stands up smoothly and steps toward the window'.
For large compositional shifts between frames, set duration to 8 seconds or longer to allow natural deceleration and acceleration.
To generate looping clips, set the exact same image URL in both slots and instruct subtle cyclic movement in prompt.
Highlight audio shifts across the motion in the prompt, such as 'starts with subtle rustling and ends on a solid closing thud'.

Notes

FLUX 3 First Last Frame to Video requires an image_urls array containing exactly two public image URLs in chronological sequence.
Frame 0 is locked to the opening still, while the final frame is locked to the end still, with intermediate frames synthesized.
After submission via API, record the returned task_id to poll progress and download the completed asset.
Rendered videos are output in standard MP4 format and can be dropped directly into timelines to link neighboring shots.

Related Endpoints

Gemini Omni 1.1 Flash Text-to-Video preview

Gemini Omni 1.1 Flash Text-to-VideoGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

Gemini Omni 1.1 Flash Image-to-Video preview

Gemini Omni 1.1 Flash Image-to-VideoGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

Gemini Omni 1.1 Flash Reference-to-Video preview

Gemini Omni 1.1 Flash Reference-to-VideoGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

Gemini Omni 1.1 Flash Video Edit preview

Gemini Omni 1.1 Flash Video EditGoogle Gemini Omni 1.1 Flash video model with native audio, text-to-video, first/last-frame image-to-video, reference-image video, and 3–10 second video edit at 360p to 4K.

MiniMax H3 Max Turbo Text to Video preview

MiniMax H3 Max Turbo Text to VideoMiniMax H3 Max Turbo Text to Video turns text into coherent 5–15 second videos with synchronized native stereo audio and fast generation. It follows physical motion while rendering subtle changes in light and shadow.

MiniMax H3 Max Turbo Image to Video preview

MiniMax H3 Max Turbo Image to VideoMiniMax H3 Max Turbo Text to Video turns text into coherent 5–15 second videos with synchronized native stereo audio and fast generation. It follows physical motion while rendering subtle changes in light and shadow.

MiniMax H3 Max Text to VideoMiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios.

MiniMax H3 Max Image to VideoMiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios.

MiniMax H3 Max Reference to Video preview

MiniMax H3 Max Reference to VideoMiniMax H3 Max Text to Video turns text prompts into cinematic video with native stereo audio, faster-than-real-time generation, and precise instruction following. It renders complex camera work and physical motion with fine detail across multiple resolutions and common aspect ratios.

Wan 3.0 Prime Text-to-VideoAlibaba Wan 3.0 Prime video model with text-to-video, image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds with optional synchronized audio.

Wan 3.0 Prime Image-to-VideoAlibaba Wan 3.0 Prime video model with text-to-video, image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds with optional synchronized audio.

Wan 3.0 Prime Reference-to-Video preview

Wan 3.0 Prime Reference-to-VideoAlibaba Wan 3.0 Prime video model with text-to-video, image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds with optional synchronized audio.

Wan 3.0 Text-to-VideoAlibaba Tongyi Lab's Wan 3.0 multimodal video model with separate text-to-video, first/last-frame image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds.

Wan 3.0 Image-to-VideoAlibaba Tongyi Lab's Wan 3.0 multimodal video model with separate text-to-video, first/last-frame image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds.

Wan 3.0 Reference-to-VideoAlibaba Tongyi Lab's Wan 3.0 multimodal video model with separate text-to-video, first/last-frame image-to-video, and reference-to-video workflows at 480p, 720p, or 1080p for 2 to 30 seconds.

FLUX 3 Text to VideoBlack Forest Labs FLUX 3 video model with text-to-video, image-to-video, first-last-frame, extend-video, and keyframes workflows at 720p or 1080p for 5 to 20 seconds, with optional native audio.

FLUX 3 First Last Frame to Video API frequently asked questions

What is the FLUX 3 First Last Frame to Video API?

FLUX 3 First Last Frame to Video is a Black Forest Labs model for synthesizing continuous video transitions between two designated still frames. Using user-supplied start and end stills as fixed composition targets, it generates 5 to 20 second video clips up to 1080p resolution with optional synchronized native audio. Built with deep motion interpolation capabilities, it preserves subject identity, spatial perspective, and surface textures while computing physically coherent camera trajectories and movement between both poles. You can call it programmatically or try it from the playground above.

How are images ordered in FLUX 3 First Last Frame to Video?

The image_urls array must contain exactly two items in chronological sequence: index 0 serves as the opening frame, and index 1 serves as the closing destination frame. The model generates forward motion from the former to the latter.

How should prompts guide the transition between frames?

Prompts should focus on how the subject moves and how the camera transitions between states. For example, 'subject stands up smoothly and steps to the desk while the camera dollies right' gives the interpolation engine explicit guidance for plausible motion paths.

How do I create seamless looping animations with FLUX 3 First Last Frame to Video?

Supply the exact same image URL as both index 0 and index 1 in image_urls, and describe a continuous cyclic action in prompt (such as a 360-degree orbit or subtle atmospheric motion). The clip will conclude seamlessly where it began.

Does FLUX 3 First Last Frame to Video generate transition audio?

Yes. When sound is true, the model evaluates visual changes across both frames alongside prompt instructions to synthesize matching dynamic sound effects and ambient progression.

What duration is recommended for wide differences between frames?

If there is significant displacement or a complex posture shift between the start and end images, setting duration to 8–15 seconds gives the physics engine sufficient time to render natural acceleration and deceleration.

Which endpoint should I use to direct more than two keyframe points?

This endpoint specializes in dual-anchor interpolation. To position up to 10 keyframes along a 24 fps timeline, use the FLUX 3 Keyframes to Video endpoint for comprehensive multi-shot storyboard directing.