Nowadays, the performance of domestic and foreign big models such as DeepSeek is becoming more and more powerful, and the industry generally believes that AI applications will usher in explosive growth in 2025. However, for enterprises, even with powerful big models, they still face the problems of unclear application scenarios and uncertain application forms. How to practically land the big model technology into industry applications and develop truly valuable AI products has been the focus of the industry's attention in the past year, and is also a problem that many enterprises continue to explore.

Based on our long experience with the beanbag modeling service, we note thatvolcanic arkRecently launched the "Big Model Application Lab", whose core features are "easy to integrate, easy to land, more open". Simply put, Volcano Ark provides a series of selected application scenarios for enterprises and develops high-quality AI applications as industry templates, which are provided in the form of open source for enterprises in need.

Interactive bilingual video generator: a new paradigm for AI-enabled educational animation

As Agent developers, our team became interested in an application called "Interactive Bilingual Video Generator" in Volcano Ark and decided to deploy and test it. We hope to take this opportunity to explore the potential of AI animation in education.

Rapid Deployment Guide

For the reader's understanding and ease of operation, the key steps are explained in detail below.

First, the specific code repository needs to be cloned:

# Repository Download
git clone https://github.com/volcengine/ai-app-lab.git
# Go to the specific directory
cd demohouse/chat2cartoon

Next, open the .env file to configure environment variables. You need to configure the parameters related to the models for text-generated graphs, speech synthesis, video generation, and video understanding.

# Large model access point ID for script creation, subplotting, roles https://console.volcengine.com/ark/region:ark+cn-beijing/openManagement?LLM=& OpenTokenDrawer=false
LLM_ENDPOINT_ID='ep-xxx'
# Visual Understanding Large Model Access Point ID for final video movie interaction
VLM_ENDPOINT_ID='ep-2025xxx'
# Volcano Engine TOS storage bucket name for storing model artifacts https://console.volcengine.com/tos/bucket/
TOS_BUCKET='chat2'
# Voice Technology API Access Key https://console.volcengine.com/speech/service/
TTS_ACCESS_KEY='7naxxx'
# Voice Technology API Resource ID https://console.volcengine.com/speech/service/
TTS_API_RESOURCE_ID='volc.service_type.10029'
# Voice Technology App Key https://console.volcengine.com/speech/service/
TTS_APP_KEY='113xxx'
# Generate Video Large Model Access Point ID (only Doubao-Video Generation Model is supported for now)
CGT_ENDPOINT_ID='ep-20250306153842-pg2b4'
# Volcano Ark API Key for Ark model access point reasoning when doing authentication https://console.volcengine.com/ark/region:ark+cn-beijing/apiKey
ARK_API_KEY='99831b24-55xxxx'
# Volcano Engine Account Access Key for accessing TOS API and uploading model products https://console.volcengine.com/iam/keymanage/
VOLC_ACCESSKEY='AKLTYxxxx'
# Volcano Engine account Secret Key, used to access TOS API, upload model products https://console.volcengine.com/iam/keymanage/
VOLC_SECRETKEY='Tmprexxxx'

Volcano Ark service opening and configuration

First of all, you need to open the relevant services of Volcano Ark (all kinds of AI models are provided on this platform). After logging in to Volcano Ark, find and click "Open Management" in the lower left corner of the page, and open the services of big language model and visual big model respectively.

After opening the modeling service, you need to create the access point, which is the actual model to be used. Click "Online Reasoning" on the left side, then click "Customize Reasoning Access Point" to create an inference access point.

Fill in the information according to the page prompts, add the specific model required and then confirm the access.

After successful creation, copy the access point ID.

Specific model choices can be adjusted according to actual needs and preferences. In this test, we chose the following models:

1. LLM_ENDPOINT_ID option Doubao-1.5-pro-32k
2. VLM_ENDPOINT_ID option Doubao-vision-pro-32k
3. CGT_ENDPOINT_ID option Doubao-Video Generation-Seaweed

To get the API Key for these models (i.e. ARK_API_KEYIf you want to create a new API Key, you can manage it in the bottom left corner of the page. If you need to create a new API Key, you can manage it in the bottom left corner of the page.

TOS Storage Bucket Configuration

Click into the created TOS storage bucket to configure cross-domain access.

Please adjust the specific parameter configuration according to the actual application scenario. The parameter configurations provided in this article are only examples for reference (please be careful when configuring the production environment).

Volcano Engine Access Control

Next, go to the Volcano Engine's Access Control page:

https://console.volcengine.com/iam/keymanage/

Gets the Access Key and Secret Key of the Volcano Engine for accessing the TOS API.

corresponding to .env in the file VOLC_ACCESSKEY cap (a poem) VOLC_SECRETKEY Parameters.

Object Storage Configuration

The TOS API is used to upload model-generated files. Go to the Object Storage page:

https://console.volcengine.com/tos

Click "Bucket List", then click "Create Bucket", fill in the relevant information to create a storage bucket. In this example, the name of the created bucket is chat2Therefore .env Papers TOS_BUCKET The parameter should be set to chat2The

Voice technology configuration

Finally, the voice technology section is configured. Visit the Volcano Engine speech technology platform:

https://console.volcengine.com/speech/app

Create an application and select the "Large Model Speech Synthesis" and "Streaming Speech Recognition Large Model" services.

Once created, click on any menu on the left to find the APP ID and Access Token below.

According to the official Volcano Engine documentation.

TTS_ACCESS_KEY corresponding to the Access Token.

TTS_APP_KEY Corresponds to the APP ID.

https://www.volcengine.com/docs/6561/1329505

Up to this point..env The configuration of the files has been completed. Next, you need to install the project dependencies and run the program.

back-end operation

# Enter the backend
cd backend
# install poetry
pip install poetry==1.6.1
# install dependencies with poetry
poetry install
# backend startup!
poetry run python index.py

If the run is successful, the terminal will display output similar to the following message.

front-end operation

# Go to frontend
cd frontend
# install pnpm
npm install -g pnpm@8
# Install dependencies with pnpm
pnpm install
# Copy environment variable .env file
cp ... /.env . /.env .
# front-end startup!
pnpm dev

If the run is successful, the terminal will display output similar to the following message.

Once you have completed the above steps, you can visit in your browser the http://localhost:8080/ Start using the Interactive Bilingual Video Generator.

Project Architecture and Test Results

The overall process architecture of the project is shown below:

Test results show that "Interactive Bilingual Video Generator" supports users to generate minute videos with one click, which is extremely easy and efficient to operate. Users do not need to set up cumbersome parameters, just enter the video requirements, you can quickly generate a long video work that meets the requirements, thus greatly improving the efficiency of creation.

The generated videos are of high quality, with clear and smooth graphics and a coherent and natural storyline. In addition, the app supports interactive Q&A with users about the video content.

Applying open source: a critical step in getting big models off the ground

surname Cong Coze The templated application of the platform to the launch of the Volcano Ark AI Application Open Source Lab not only represents the extension of the solution from low-code to high-code, but also marks the evolution of the application scenarios from generality to deep customization.

In the wave of big model technology application, the strategic significance of application open source is even beyond the model open source itself. It is true that a powerful model is the engine of AI application, but how to efficiently integrate the model capability into actual business scenarios is the key to promote the landing of AI application and ultimately improve business capability.

Volcano Engine Open Source AI Lab provides open source, high-code SDKs and prototype AI applications, which precisely fill in the "last kilometer" for the landing of AI applications. Open-source AI applications provide the best solution for enterprises to start quickly.

Although many enterprises recognize the huge potential of large models and understand how to apply them to their business scenarios from a theoretical level, they still face many obstacles in actual operation. The emergence of open-source AI prototype applications allows enterprise developers not to start from scratch to figure out the complex model docking and application development process, and can quickly get started, quickly learn and build and expand AI applications that meet their business needs, thus significantly reducing the cost of trial-and-error, time costs and labor costs.

For the majority of AI technology enthusiasts and developers, when they first get involved in the field of AI application development, they will often come into contact with highly encapsulated frameworks with a high degree of abstraction, such as LangChain. LangChain framework in skilled mastery, can indeed significantly improve the development efficiency, but its large number of syntactic sugar and abstract concepts, but also to the beginner to bring a higher learning threshold. In contrast, the Python SDK Arkitect provided by Volcano Engine is much easier to get started, and its toolchain and development process are more intuitive. In addition, the official Demo also provides a detailed technical architecture diagram and implementation details, which is convenient for developers to deeply understand.

The launch of the Volcano Ark AI Application Lab undoubtedly provides a powerful AI application development platform for enterprises and developers. It is especially commendable that its open source strategy has lowered the threshold of AI application development and accelerated the landing process of big model technology in various industries. With the emergence of more open source applications, we have reason to believe that AI technology will be truly integrated into thousands of industries and unleash greater potential.