You need to enable JavaScript to run this app.
Lake AI Service

Lake AI Service

Copy page
Download PDF
Quick start
Quick start
Copy page
Download PDF
Quick start

Try out data processing operators within a few minutes.

Experience LAS operator capabilities

LAS provides the online experience feature and offers a certain amount of free quota. No configuration is required; you can quickly process sample documents, audio, video, and other files through the online experience portal to try out the multimodal data processing capabilities provided by LAS.

warning

Currently, LAS provides a certain amount of free quota for the online experience. Any usage beyond this quota will be billed according to the operator's billing items.

Online experience portal

After logging in to the LAS Console , you can view the card information for each operator on the Operator Square page. Online operators support quick trials; you can click the "Online Experience" portal on the corresponding card to conveniently try out the operator's multimodal data processing capabilities.
Image

Online experience operation demonstration

The following is an example of using the online experience feature to try out the capabilities of LAS operators.

  • LAS provides multiple sample video files. You can also delete the sample files and upload your own files manually.
  • During the online experience, you can adjust the operator's processing parameters as needed.

The free quota limits for various operators supported by LAS in the online experience are as follows. Please monitor your usage during the trial to avoid incurring charges for exceeding the quota.

Operator

Free experience quota

Over-quota billing

Enhanced video content understanding (Doubao series)

500,000 tokens

  • During the online experience, if you exceed the free quota, the excess usage will be billed according to the operator's billing items.
  • For the billing items and billing logic of each operator, see Large model billing.

PDF document parsing (Doubao)

50 pages

Speech-to-text (Doubao Speech ASR)

5 minutes

Speech-to-text (Doubao audio file recognition) enhanced version

5 minutes

Image-text embedding (Doubao series models)

500,000 tokens

Quickly invoke LAS operators

1. Obtain and configure LAS API Key

  1. After logging in to the LAS console, go to the Resource Management > API Key Management page and click Create API Key in the upper left corner. Create the API Key as prompted by the interface.
    Image
  2. Configure environment variables. Run the following command in the terminal to set the API Key as an environment variable.
    export LAS_API_KEY="your_api_key_here"
    

2. Obtain the Base URL

The Base URL depends on the region of the LAS service you are using. For operator capability trials, it is recommended to use the Beijing region Base URL: https://operator.las.cn-beijing.volces.com/api/v1.

3. Send an API request

The following example uses the Multimodal deep reasoning (Doubao-seed-1.8) operator for a quick trial. Run the following curl command on your development machine to quickly see sample results of LAS operators for image understanding and processing.

Input

Output preview

  • Test image URL: https://ark-project.tos-cn-beijing.volces.com/images/view.jpeg
  • Test image details:
    Image

This is an outdoor scene filled with a sense of natural tranquility:

  • Foreground: The expansive deep blue lake surface ripples gently. A bright orange single-person kayak floats on the water, with one person sitting in it, raising a black paddle, seemingly rowing. The water clearly reflects the kayak and the person's silhouette. - Midground: Across the lake is a dense, dark green coniferous forest, with a light mist lingering among the trees, adding a touch of haziness to the woods. - Background: Behind the forest are towering, continuous snow-capped mountains. The mountains are covered in ice and snow, and their rugged contours are clearly visible under the interplay of light and shadow, displaying a majestic and grand atmosphere. - Sky: Above is a clear, pale blue sky, dotted with a few wisps of thin clouds. The overall atmosphere is open and serene, conveying the leisure and vastness of outdoor nature. "reasoning_content":"The user now needs to describe the content of this image. First, looking at the overall scene: the foreground is a calm, deep blue lake. On the water is an orange kayak with one person in it, raising a black paddle, seemingly rowing. The water reflects both the person and the kayak." Then, in the middle part, across the lake is a dense, dark green coniferous forest. Above the forest is a light mist, creating a tranquil atmosphere. The background is a towering snow-capped mountain, with the mountain covered in snow and its rugged contours standing out under the blue sky, creating a majestic scene. The sky is a pale blue, with a few wisps of thin white clouds drifting across. The overall atmosphere is tranquil and spacious, evoking a relaxed sense of outdoor nature. Now organize these elements clearly and describe them in layers. First, focus on the main subject:
  1. Water surface and kayak: The foreground features a wide, calm, deep blue lake with gentle ripples on the water. A bright orange single-person kayak floats on the lake, with a person seated in it, raising a black paddle as if in the act of rowing. The water reflects the silhouette of both the kayak and the person. 2. Midground: On the far side of the lake is a dense, dark coniferous forest (appearing to be firs, pines, or similar trees), with a faint mist among the trees, giving the forest a hazy appearance. 3. Background: Behind the forest is a range of majestic snow-capped mountains. The mountains are covered in ice and snow, their contours clearly defined, and the interplay of light and shadow reveals the texture of rocks and snow, making them appear grand and imposing. 4. Sky: Above is a clear, pale blue sky adorned with a few wispy, thin white clouds. The weather is bright and clear, giving the entire scene a tranquil and open atmosphere, filled with the serenity and grandeur of nature. Now compose these into a smooth, well-structured description to make the content clear and well-organized.

The operator invocation code is as follows.

curl --location "https://operator.las.cn-beijing.volces.com/api/v1/chat/completions" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer $LAS_API_KEY" \
--data '{
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://ark-project.tos-cn-beijing.volces.com/images/view.jpeg"
                    }
                },
                {
                    "type": "text",
                    "text": "图片里面有什么"
                }
            ]
        }
    ],
    "model": "doubao-seed-1-8-251228"
}'

Expand your knowledge
  • Business migration: Compatible with OpenAI SDK: OpenAI SDK compatibility note
  • Operator categories: LAS provides both offline and online data processing operators.
    • Online operators: Invoked directly via the operator API.
    • Offline operators: Invoked using the Daft engine. LAS provides the VeDaft engine, which integrates the SDK for offline operator invocation, enabling more efficient scheduling and execution of batch data processing tasks.
  • When calling the official online operator OpenAPI via HTTP, there are two invocation modes: synchronous and asynchronous.
    • Synchronous invocation allows you to request directly and obtain the data processing result in the response.
    • Asynchronous invocation requires submitting a job first, then using the returned job ID to query the job and obtain the business result.

Next steps

After experiencing operator capabilities, you can use test data for end-to-end hands-on practice. For details, see:

Last updated: 2026.05.22 13:58:52