Clip Generation API (Reka Clip)

For chat over text, images and short videos, use the OpenAI-compatible Chat Completions API.

Usage

There are two ways to generate clips from your videos:

Option 1: 2‑step workflow
- Submit a job with POST /v1/clips
- Poll job status with GET /v1/clips/{id} until completed
Option 2: Single streaming request
- Submit a job to POST /v1/clips with stream=true to receive real‑time progress updates (SSE)

Option 1: Submit and poll

Submit a generation job

Endpoint

POST /v1/clips

Headers

Content-Type: application/json
X-Api-Key: YOUR_API_KEY

Request body

video_urls (array[string], required): URLs of input videos

Must contain at least one URL
Supports YouTube, Twitch, direct MP4 links

prompt (string, optional): Description of the clip to generate

Defaults to Create an engaging short video highlighting the best moments

generation_config (object, optional):

Field	Type	Default	Notes
`template`	`moments \| compilation`	moments	Generation style
`num_generations`	integer (1–3)	1	Number of clips to generate
`min_duration_seconds`	integer ≥ 0	0	Minimum clip length
`max_duration_seconds`	integer (1–600)	90	Maximum clip length
`source_start_time`	integer	—	Start time for source video (seconds)
`source_end_time`	integer	—	End time for source video (seconds)

rendering_config (object, optional):

Field	Type	Default	Notes
`show_watermark`	boolean	false	Display Reka watermark
`subtitles`	boolean	true	Enable subtitles
`aspect_ratio`	`9:16 \| 9:16-split \| 16:9 \| 4:5 \| 1:1`	9:16	Output aspect ratio. *Note: `9:16-split` means the clip will include split screen reframing where optimal.
`resolution`	`240 \| 360 \| 480 \| 720 \| 1080`	720	1080p may incur HD processing fee
`caption_style`	object	—	Optional subtitle styling configuration (see below)

Resolution and HD processing: The API tries to deliver the requested resolution. If it delivers a lower one, you are not charged the HD fee.

caption_style (object, optional):

Field	Type	Default	Notes
`desired_font_size`	number (60–120)	120	Subtitle font size
`text_transform`	`initial \| uppercase \| lowercase`	uppercase	Text casing
`text_color`	hex string	#FFFFFF
`highlight_color`	hex string	#FF7E4F
`stroke_color`	hex string	#000000
`position`	`top \| middle \| bottom`	bottom	Subtitle position
`font_family`	`BebasNeue \| Bangers \| RobotoCondensed \| Lato \| CaptionFont`	BebasNeue

Validation: The API ignores unsupported keys and invalid or out-of-range values.

Request example

Bash

curl -X POST "https://vision-agent.api.reka.ai/v1/clips" \
  -H "X-Api-Key: $REKA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_urls": ["https://www.youtube.com/watch?v=U0Cd8jbj5zg"],
    "prompt": "Create an engaging short video highlighting the best moments",
    "generation_config": {
      "template": "moments",
      "num_generations": 3,
      "min_duration_seconds": 0,
      "max_duration_seconds": 60
    },
    "rendering_config": {
      "subtitles": true,
      "aspect_ratio": "9:16",
      "resolution": 720,
      "caption_style": {
            "desired_font_size": 120,
            "text_transform": "uppercase",
            "text_color": "#FFFFFF",
            "highlight_color": "#FF7E4F",
            "stroke_color": "#000000",
            "position": "middle",
            "font_family": "BebasNeue"
          }
    }
  }'

Python

import requests, json
BASE_URL = "https://vision-agent.api.reka.ai"
API_KEY = "YOUR_API_KEY"  # or use an environment variable
endpoint_url = f"{BASE_URL}/v1/clips"
payload = {
    "video_urls": ["https://www.youtube.com/watch?v=U0Cd8jbj5zg"],
    "prompt": "Create an engaging short video highlighting the best moments",
    "generation_config": {
        "template": "moments",
        "num_generations": 3,
        "min_duration_seconds": 0,
        "max_duration_seconds": 60,
    },
    "rendering_config": {
        "subtitles": True,
        "aspect_ratio": "9:16",
        "resolution": 720,
        "caption_style": {
              "desired_font_size": 120,
              "text_transform": "uppercase",
              "text_color": "#FFFFFF",
              "highlight_color": "#FF7E4F",
              "stroke_color": "#000000",
              "position": "middle",
              "font_family": "BebasNeue"
            }
    },
}
headers = {
    "Content-Type": "application/json",
    "X-Api-Key": API_KEY,
}
response = requests.post(endpoint_url, headers=headers, json=payload)
response.raise_for_status()
result = response.json()
print(json.dumps(result, indent=2))
generation_id = result["id"]

Response example (job accepted)

{
  "id": "156b59b3-82a2-4619-99c5-954d80d3e254",
  "status": "queued",
  "created_at": "2026-01-17T19:23:20.878691+00:00",
  "updated_at": "2026-01-17T19:23:20.878691+00:00",
  "generation_config": {
    "template": "moments",
    "num_generations": 3,
    "min_duration_seconds": 0,
    "max_duration_seconds": 60
  },
  "rendering_config": {
    "subtitles": true,
    "aspect_ratio": "9:16",
    "resolution": 720,
    "caption_style": {
          "desired_font_size": 120,
          "text_transform": "uppercase",
          "text_color": "#FFFFFF",
          "highlight_color": "#FF7E4F",
          "stroke_color": "#000000",
          "position": "middle",
          "font_family": "BebasNeue"
        }
  },
  "video_urls": [
    "https://www.youtube.com/watch?v=1vM3Xsiwsao"
  ],
  "prompt": "Create an engaging short video highlighting the best moments"
}

A newly submitted job starts with status set to queued.

Save the id. You need it to check generation status.

Check generation status

GET /v1/clips/{id}

Returns job metadata and, when status is completed, an output array with generated clips.

Request example

Bash

GEN_ID="YOUR_GENERATION_ID"
curl -X GET "https://vision-agent.api.reka.ai/v1/clips/${GEN_ID}" \
  -H "X-Api-Key: $REKA_API_KEY" \
  -H "Content-Type: application/json" | jq .

Python

url = f"{BASE_URL}/v1/clips/{generation_id}"
headers = {"X-Api-Key": API_KEY}
resp = requests.get(url, headers=headers)
resp.raise_for_status()
status = resp.json()
print(json.dumps(status, indent=2))

Response example (processing)

{
  "id": "be22635b-8d25-4c7a-b885-b1dcba82973d",
  "status": "processing",
  "created_at": "2026-01-18T19:10:13.36329+00:00",
  "updated_at": "2026-01-18T19:10:13.36329+00:00",
  "generation_config": {
    "template": "moments",
    "num_generations": 3,
    "min_duration_seconds": 0,
    "max_duration_seconds": 60
  },
  "rendering_config": {
    "subtitles": true,
    "aspect_ratio": "9:16"
  },
  "video_urls": [
    "https://www.youtube.com/watch?v=x6-vh91YnuY"
  ],
  "prompt": "Create an engaging short video highlighting the best moments",
  "output": [],
  "error_message": null
}

The status field shows the job’s current stage, processing in this example.

Job lifecycle

A job moves through these states:

Once the status is completed, the response includes the generated clips:

Response example (completed)

{
  "id": "156b59b3-82a2-4619-99c5-954d80d3e254",
  "status": "completed",
  "created_at": "2026-01-17T19:23:20.878691+00:00",
  "updated_at": "2026-01-17T19:27:03.650333+00:00",
  "generation_config": {
    "template": "moments",
    "num_generations": 3,
    "min_duration_seconds": 0,
    "max_duration_seconds": 60
  },
  "rendering_config": {
    "subtitles": true,
    "aspect_ratio": "9:16"
  },
  "video_urls": [
    "https://www.youtube.com/watch?v=1vM3Xsiwsao"
  ],
  "prompt": "Create an engaging short video highlighting the best moments",
  "output": [
    {
      "title": "Marcelo's Wild Family Party Stories",
      "video_url": "https://guardian-user-uploads-prod.cdn.reka.ai/org-a10e5eec-1028-49fa-9144-ef676e3328e1/reels/2ab608e5b0d7704b521aae004d1d7b2c",
      "caption": "When your cousins make you do the most unhinged things at family parties 😭💀 Marcelo's comedy is unmatched!",
      "hashtags": [
        "#MarceloHernandez",
        "#StandUp",
        "#Comedy",
        "#FamilyParty",
        "#Cousins",
        "#LiveComedy",
        "#Funny",
        "#Relatable",
        "#Netflix",
        "#AmericanBoy"
      ],
      "ai_score": 82
    },
    {
      "title": "The Ultimate Test Fail Story",
      "video_url": "https://guardian-user-uploads-prod.cdn.reka.ai/org-a10e5eec-1028-49fa-9144-ef676e3328e1/reels/53b91df6844880cf225d617dc0e2b282",
      "caption": "POV: You feel like a genius on a test but forget to write your NAME 💀 Marcelo's story is too relatable!",
      "hashtags": [
        "#MarceloHernandez",
        "#TestFail",
        "#Epic",
        "#Relatable",
        "#Comedy",
        "#StoryTime",
        "#Funny",
        "#Netflix",
        "#AmericanBoy",
        "#LateNight"
      ],
      "ai_score": 88
    },
    {
      "title": "Marcelo's Viral Sebastian Maniscalco Impression Tutorial",
      "video_url": "https://guardian-user-uploads-prod.cdn.reka.ai/org-a10e5eec-1028-49fa-9144-ef676e3328e1/reels/702d8d67dff7cae2c2b06dedead0de05",
      "caption": "When someone asks you to break down your viral SNL impression 😂 Marcelo's Sebastian is TOO accurate!",
      "hashtags": [
        "#MarceloHernandez",
        "#SebastianManiscalco",
        "#SNLImpression",
        "#Viral",
        "#Tutorial",
        "#Comedy",
        "#Impression",
        "#BehindTheScenes",
        "#TalkShow",
        "#Perfect"
      ],
      "ai_score": 92
    }
  ],
  "error_message": null
}

Option 2: Streaming in one request

Set stream: true to receive Server-Sent Events (SSE) with progress updates until completion.

Request example

Bash

curl -N -X POST "https://vision-agent.api.reka.ai/v1/clips" \
  -H "X-Api-Key: $REKA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "video_urls": [
      "https://demo-videos-bucket-reka.s3.eu-west-2.amazonaws.com/evYsMdBrFXc.mp4"
    ],
    "prompt": "Create an engaging short video highlighting the best moments",
    "generation_config": {
      "template": "moments",
      "num_generations": 1
    },
    "rendering_config": {
      "subtitles": true,
      "aspect_ratio": "9:16",
      "resolution": 720,
      "caption_style": {
        "desired_font_size": 120,
        "text_transform": "uppercase",
        "text_color": "#FFFFFF",
        "highlight_color": "#FF7E4F",
        "stroke_color": "#000000",
        "position": "middle",
        "font_family": "BebasNeue"
      }
    },
    "stream": true
  }'

Python

import requests, json
from typing import Generator, Dict, Any
endpoint_url = f"{BASE_URL}/v1/clips"
payload = {
    "video_urls": ["https://demo-videos-bucket-reka.s3.eu-west-2.amazonaws.com/evYsMdBrFXc.mp4"],
    "prompt": "Create an engaging short video highlighting the best moments",
    "generation_config": {
        "template": "moments",
        "num_generations": 1,
        "min_duration_seconds": 0,
        "max_duration_seconds": 90,
    },
    "rendering_config": {
        "subtitles": True,
        "aspect_ratio": "9:16",
    "resolution": 1080,
    "caption_style": {
          "desired_font_size": 120,
          "text_transform": "uppercase",
          "text_color": "#FFFFFF",
          "highlight_color": "#FF7E4F",
          "stroke_color": "#000000",
          "position": "middle",
          "font_family": "BebasNeue"
        }
    },
    "stream": True,
}
headers = {
    "Content-Type": "application/json",
    "X-Api-Key": API_KEY,
}
def stream_events(response) -> Generator[Dict[str, Any], None, None]:
    last_data = None
    for line in response.iter_lines():
        if not line:
            continue
        decoded = line.decode("utf-8")
        if decoded.startswith("data: "):
            try:
                data = json.loads(decoded[6:])
                if data != last_data:
                    yield data
                last_data = data
            except json.JSONDecodeError:
                pass
with requests.post(endpoint_url, headers=headers, json=payload, stream=True) as r:
    r.raise_for_status()
    for event in stream_events(r):
        print(json.dumps(event, indent=2))

Streaming event example

{
  "id": "ada0bc6f-3540-4bb2-baa0-308e9fcf1d06",
  "status": "preprocessing",
  "created_at": "2026-01-18T19:28:49.839838+00:00",
  "updated_at": "2026-01-18T19:29:33.202844+00:00",
  "generation_config": {
    "template": "moments",
    "num_generations": 1,
    "min_duration_seconds": 0,
    "max_duration_seconds": 90
  },
  "rendering_config": {
    "subtitles": true,
    "aspect_ratio": "9:16"
  },
  "video_urls": [
    "https://www.youtube.com/watch?v=9JbE_HtiewI"
  ],
  "prompt": "Create an engaging short video highlighting the best moments"
}

The status field shows the job’s current stage, preprocessing in this example. See Job Lifecycle for the full progression.

Error response example

{
  "id": "11add419-fc2e-4491-98fd-e0b362103ff1",
  "status": "failed",
  "created_at": "2026-01-18T19:23:05.649501+00:00",
  "updated_at": "2026-01-18T19:24:16.798044+00:00",
  "generation_config": {
    "template": "moments",
    "num_generations": 1,
    "min_duration_seconds": 0,
    "max_duration_seconds": 90
  },
  "rendering_config": {
    "subtitles": true,
    "aspect_ratio": "9:16"
  },
  "video_urls": [
    "https://www.youtube.com/watch?v=x6-vh91YnuY"
  ],
  "prompt": "Create an engaging short video highlighting the best moments",
  "error_message": "This preview of Reka Clip is limited to videos with a maximum duration of 2 hours.\nPlease contact us if you need support for longer videos!"
}

If a generation fails, the job enters a terminal failed state and includes an error_message describing the cause.

Resources

Code examples in Python, Node.js and C# are on Reka AI’s Github.

For an example of Reka Clip in n8n, see the n8n template.

To try this API without code, use Reka Clip. Paste a YouTube or Twitch link and it generates clips.