🎉 New releases: Marengo-2.6 and Pegasus-1
Read about Marengo-2.6
Read about Pegasus-1
🎉 Marengo 2.7 has been released!
Read more >
logo
Product
Research
Developers
Blog
Log inTry on Playground
logo
Product
Research
Developers
Blog
Try on Playground
Watch demo

Explore video content using language

Understand your videos with groundbreaking AI. Extract more from videos than ever before.

Try on Playground

Find any moment, when you need it

Unlock the wealth of data stored in any video. Twelve Labs can index petabytes of video content and make it semantically searchable with everyday language

Try now

Find any moment, when you need it

Unlock the wealth of data stored in any video. Twelve Labs can index petabytes of video content and make it semantically searchable with everyday language

Our Strength

Video understanding that can generate text for any video

Generate

Explore your content by leveraging our rich embeddings. Generate any open-ended text formats tailored to your needs, from Q&A and creative suggestions to content analysis.

Summarize

Synthesize content to create concise and efficient descriptions, highlights, and chapters useful for content management, exploration, and understanding.

Gist

Create topic categories, video titles, and hashtags helpful for marketing, advertising, and ideation use cases.

Pegasus
Audio and Video

Not audio or video. Pegasus naturally understands videos in their totality to understand content like you do.

Flexible

Use our out of the box templates to kickstart your video understanding or use custom prompts specific to your use case.

Fast

Quickly index once, generate text prompts in seconds.

Pushing the boundaries of video understanding — Pegasus 1

USE CASES

Generate any text about your video through prompting

Generate accurate and insightful text about your videos through prompting, be it a summary, long and detailed reports, title suggestions, highlights, or chapters.
Media and Entertainment

Deeply understand your video content like never before

Advertising and Marketing

Get more out of our campaigns

Security and Surveillance

 Empower investigations and after-incident reporting

flexibility

Deploy with flexibility

Whether you prefer native cloud, private cloud,
or on-premise, our APIs can run in any environment.
See pricing
SDK

We're developers that build for developers

Hit the ground running for any project using our dev friendly SDK. Available now in Python, and JS with Golang and Java coming soon.
Try now
Python
Javascript
Golang
Coming soon
Java
Coming soon

Python

from

import os

‍

client = TwelveLabs(<"YOUR_API_KEY">)

‍

# Create new Index

index = client.index.create(

        name="My First Index",

        engines=[

            {

                "name": "marengo2.6",

                "options": ["visual", "conversation", "text_in_video"],

            },

        ],

    )

‍

# Create new Task on Index (Upload video)

video_path = os.path.join(os.path.dirname(__file__), "<YOUR_FILE_PATH>")

task = client.task.create(index_id=index.id, file=video_path, language="en")

‍

# Wait for indexing to finish

task.wait_for_done()

‍

# Search from your index

query = "An artist climbing up the ladder that he painted."

result = client.search.query(index.id, query, ["visual", "conversation"])

print(result)

‍

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35

from twelvelabs import TwelveLabs
import os

client = TwelveLabs(<"YOUR_API_KEY">)

# Create new Index
index = client.index.create(
      name="My First Index",
      engines=[
          {
              "name": "pegasus1",
              "options": ["visual", "conversation"],
          },
      ],
   )

# Create new Task on Index (Upload video)
video_path = os.path.join(os.path.dirname(__file__), "<YOUR_FILE_PATH>")
task = client.task.create(index_id=index.id, file=video_path, language="en")

# Wait for indexing to finish
task.wait_for_done()
‍
# Generate Open-ended text from your video
prompt = "Give me an analysis of how the advertisement uses facial expressions to convey its message."
res = client.generate.text(task.video_id, prompt)
print(res.data)
‍

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
‍

The advertisement effectively uses facial expressions to convey its message. The child brushing their teeth is shown with a big smile, indicating that dental health is enjoyable and can bring joy. The child's dancing and pointing towards the camera further emphasizes the enthusiasm and excitement for brushing.The woman holding the child on the sofa also displays a warm and caring expression, indicating the importance of parental involvement in dental hygiene.

Python

from

import os

‍

client = TwelveLabs(<"YOUR_API_KEY">)

‍

# Create new Index

index = client.index.create(

        name="My First Index",

        engines=[

            {

                "name": "marengo2.6",

                "options": ["visual", "conversation", "text_in_video"],

            },

        ],

    )

‍

# Create new Task on Index (Upload video)

video_path = os.path.join(os.path.dirname(__file__), "<YOUR_FILE_PATH>")

task = client.task.create(index_id=index.id, file=video_path, language="en")

‍

# Wait for indexing to finish

task.wait_for_done()

‍

# Search from your index

query = "An artist climbing up the ladder that he painted."

result = client.search.query(index.id, query, ["visual", "conversation"])

print(result)

‍

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

import { TwelveLabs } from 'twelvelabs';
import fs from 'fs';
import path from 'path';

const client = new TwelveLabs({ apiKey: '<YOUR_API_KEY>' });

// Create new Index
const index = await client.index.create({
    name: 'My First Index',
    engines:[
      {
        name: 'marengo2.6',
        options: ['visual', 'conversation', 'text_in_video'],
      },
    ]
  });

// Create new Task on Index (Upload the video)
const videoPath = path.join(__dirname, '<YOUR_FILE_PATH>');
const task = await client.task.create({
  indexId: index.id,
  file: fs.createReadStream(videoPath),
  language: 'en',
});

// Wait for indexing to finish
await task.waitForDone();

// Generate Open-ended text from your video
const prompt = "Give me an analysis of how the advertisement uses facial expressions to convey its message."
const result = await client.generate.text(task.videoId, prompt)
prconsole.log(result)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
‍

The advertisement effectively uses facial expressions to convey its message. The child brushing their teeth is shown with a big smile, indicating that dental health is enjoyable and can bring joy. The child's dancing and pointing towards the camera further emphasizes the enthusiasm and excitement for brushing.The woman holding the child on the sofa also displays a warm and caring expression, indicating the importance of parental involvement in dental hygiene.

Coming soon
Coming soon
Python
Javascript

Python

from

import os

‍

client = TwelveLabs(<"YOUR_API_KEY">)

‍

# Create new Index

index = client.index.create(

        name="My First Index",

        engines=[

            {

                "name": "marengo2.6",

                "options": ["visual", "conversation", "text_in_video"],

            },

        ],

    )

‍

# Create new Task on Index (Upload video)

video_path = os.path.join(os.path.dirname(__file__), "<YOUR_FILE_PATH>")

task = client.task.create(index_id=index.id, file=video_path, language="en")

‍

# Wait for indexing to finish

task.wait_for_done()

‍

# Search from your index

query = "An artist climbing up the ladder that he painted."

result = client.search.query(index.id, query, ["visual", "conversation"])

print(result)

‍

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

import { TwelveLabs } from 'twelvelabs';
import fs from 'fs';
import path from 'path';

const client = new TwelveLabs({ apiKey: '<YOUR_API_KEY>' });

// Create new Index
const index = await client.index.create({
    name: 'My First Index',
    engines:[
      {
        name: 'marengo2.6',
        options: ['visual', 'conversation', 'text_in_video'],
      },
    ]
  });

// Create new Task on Index (Upload the video)
const videoPath = path.join(__dirname, '<YOUR_FILE_PATH>');
const task = await client.task.create({
  indexId: index.id,
  file: fs.createReadStream(videoPath),
  language: 'en',
});

// Wait for indexing to finish
await task.waitForDone();

// Generate Open-ended text from your video
const prompt = "Give me an analysis of how the advertisement uses facial expressions to convey its message."
const result = await client.generate.text(task.videoId, prompt)
prconsole.log(result)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
‍
‍

The advertisement effectively uses facial expressions to convey its message. The child brushing their teeth is shown with a big smile, indicating that dental health is enjoyable and can bring joy. The child's dancing and pointing towards the camera further emphasizes the enthusiasm and excitement for brushing.The woman holding the child on the sofa also displays a warm and caring expression, indicating the importance of parental involvement in dental hygiene.

Python

from

import os

‍

client = TwelveLabs(<"YOUR_API_KEY">)

‍

# Create new Index

index = client.index.create(

        name="My First Index",

        engines=[

            {

                "name": "marengo2.6",

                "options": ["visual", "conversation", "text_in_video"],

            },

        ],

    )

‍

# Create new Task on Index (Upload video)

video_path = os.path.join(os.path.dirname(__file__), "<YOUR_FILE_PATH>")

task = client.task.create(index_id=index.id, file=video_path, language="en")

‍

# Wait for indexing to finish

task.wait_for_done()

‍

# Search from your index

query = "An artist climbing up the ladder that he painted."

result = client.search.query(index.id, query, ["visual", "conversation"])

print(result)

‍

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

import { TwelveLabs } from 'twelvelabs';
import fs from 'fs';
import path from 'path';

const client = new TwelveLabs({ apiKey: '<YOUR_API_KEY>' });

# Create new Index
const index = await client.index.create({
    name: 'My First Index',
    engines:[
      {
        name: 'marengo2.6',
        options: ['visual', 'conversation', 'text_in_video'],
      },
    ]
  });

# Create new Task on Index (Upload video)
const videoPath = path.join(__dirname, '<YOUR_FILE_PATH>');
const task = await client.task.create({
  indexId: index.id,
  file: fs.createReadStream(videoPath),
  language: 'en',
});

# Wait for indexing to finish
await task.waitForDone();

# Generate Open-ended text from your video
prompt = "Give me an analysis of how the advertisement uses facial expressions to convey its message."
res = await client.generate.text(task.videoId, prompt)
print(res.data)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
‍

The advertisement effectively uses facial expressions to convey its message. The child brushing their teeth is shown with a big smile, indicating that dental health is enjoyable and can bring joy. The child's dancing and pointing towards the camera further emphasizes the enthusiasm and excitement for brushing.The woman holding the child on the sofa also displays a warm and caring expression, indicating the importance of parental involvement in dental hygiene.

Copyright Twelve Labs. All rights reserved.