LangStudio overview
Due to a product strategy adjustment, PAI-LangStudio will be retired on March 9, 2027 at 16:00. After the retirement date, you will no longer be able to use LangStudio features in the PAI console, and all associated support services will also be discontinued. For more information, see PAI-LangStudio retirement and migration guide.
Overview
PAI-LangStudio is an all-in-one enterprise-grade agent development platform built on Alibaba Cloud PAI. The platform supports both low-code (visual workflow orchestration) and high-code development modes. It provides built-in components, knowledge bases, MCP integration, and tools that help you efficiently build, debug, and iterate on agent apps.
PAI-LangStudio is designed for enterprise use cases and offers multi-region deployment worldwide. Enterprise features include development/production environment isolation, fine-grained access control through Alibaba Cloud RAM, dedicated VPCs and compute resources, evaluation frameworks, and full-stack observability. These capabilities ensure that your workloads are secure, stable, and compliant. You can quickly validate POCs and publish them as production-ready, highly available services, bridging the gap from prototype to production.
Feature architecture
LangStudio covers the full lifecycle of agent app development, from rapid development to quality optimization and enterprise-grade deployment:
Rapid agent app development
The platform supports both low-code (visual workflow orchestration) and high-code development modes. With ready-to-use components and templates, unified connection configuration, and trace analysis, you can efficiently build, debug, and iterate on agent apps.
Visual workflow orchestration: Drag and connect nodes such as foundation models, Python code, and tool calls on a graphical canvas to quickly build application workflows.
High-code development: A full IDE environment that supports mainstream agent frameworks such as AgentScope and LangGraph, enabling you to implement complex business logic through code.
Application quality optimization
Agent apps commonly face issues such as hallucination, context loss, and low accuracy. The platform addresses these with multi-layered optimization: memory optimization, context management, automatic planning, multimodal knowledge bases, evaluation, and model tuning.
Multimodal knowledge base: Supports retrieval of multimodal data such as files, images, and videos. Built on enterprise-selected vector databases (Hologres, OpenSearch, Milvus, and others) to ensure full control over data assets. Automatic index updates and incremental synchronization keep knowledge up to date.
Automatic planning: Agents can autonomously break tasks into steps and invoke tools and knowledge bases to complete complex tasks.
Evaluation framework: A comprehensive application evaluation capability that drives continuous improvement through batch testing.
Enterprise capabilities
Designed for production environments, the platform provides multi-region deployment worldwide, development/production environment isolation, fine-grained access control through RAM, dedicated VPCs and compute resources, and full-stack observability to ensure application security, stability, and compliance.
Fine-grained access control: Built on the Alibaba Cloud RAM authorization system with support for resource-level permission isolation. PAI workspace RBAC authorization is also supported.
VPC network isolation: Supports deployment in dedicated VPCs with dedicated compute resources to ensure the security of data in transit and at rest.
Independent model deployment: Integrates with PAI-EAS and ModelGallery for fast, independent model deployment that ensures performance and stability.
Full-stack observability: Integrates with Alibaba Cloud Tracing Analysis and Log Service (SLS) to provide metrics monitoring, alerting, distributed tracing, and log analysis.
Benefits
Quick start with ready-to-use templates: Includes high-quality application templates (DeepResearch, RAG, NL2SQL, and more) refined through real-world internal use. These templates help you go from concept to production quickly, reducing the learning curve and time to first deployment.
Seamless PAI ecosystem integration: Covers the full agent app lifecycle, including application debugging, model tuning, trace analysis, evaluation, online deployment, and full-stack observability. This enables AI applications to move rapidly from POC validation to production deployment.
White-box pipeline with full data control: Provides a transparent pipeline for data cleansing, chunking strategies, and index building. Enterprises can customize search and vectorization approaches as needed. All sensitive data (source data, processing logic, intermediate data, and indexes) remains under complete enterprise control.
Enterprise-grade security and reliability: Ensures high availability and security for enterprise applications through VPC network isolation, role-based fine-grained access control, independent PAI-EAS deployment, and content security review.
Use cases
LangStudio offers flexible features that can be combined based on your business scenarios. The following table describes typical use cases:
|
Use case |
Description |
Key capabilities |
|
Enterprise RAG application |
Build retrieval-augmented generation applications with enterprise private knowledge bases to improve accuracy and timeliness in domain-specific Q&A. |
Multimodal knowledge base, dynamic index updates, security isolation, context management |
|
NL2SQL intelligent BI |
Query databases in natural language, auto-generate SQL, and produce reports to lower the barrier for data analysis. |
Natural language understanding, SQL generation, database connectivity, automated reporting |
|
Multimodal chat agent |
Handle multimodal inputs such as text, voice, and images for complex interaction scenarios including AI-powered customer service and content creation. |
Multimodal understanding, state management, tool calling, task planning |
|
Intelligent voice assistant |
Integrate speech synthesis and recognition to build voice interaction applications for scenarios such as in-vehicle systems and smart home devices. |
Speech recognition, speech synthesis, dialog management, tool calling |
|
Intelligent design assistant |
Combine visual generation capabilities to assist designers with creative design, image editing, and other tasks. |
Visual generation, prompt engineering, creative optimization, multi-turn interaction |
Supported regions
LangStudio is available in the following regions: China (Hangzhou), China (Shanghai), China (Beijing), China (Ulanqab), China (Shenzhen), China (Hong Kong), Singapore, Japan (Tokyo), Indonesia (Jakarta), Germany (Frankfurt), and US (Virginia).
To minimize latency, select a region that is close to your data sources or end users.
Billing
LangStudio itself does not charge additional service fees. You pay only for the underlying cloud resources you use. Costs may include Object Storage Service (OSS) for storing workspace files, Tracing Analysis for distributed tracing, Log Service (SLS) for log collection, and PAI-EAS for service deployment. For more information, see LangStudio billing.
Getting started
The typical workflow for using LangStudio is as follows:
To help you get started quickly, LangStudio provides detailed quick start guides: Create a workflow application and Create a code-based application.
