Studio Overview

What is Studio?

aiXplain Studio is your workspace for discovering, building, and deploying AI agents. Studio provides two main experiences:

• Browse - Explore and discover AI assets including models, agents, tools, and integrations
• Build - Create and deploy single agents or multi-agent systems (team agents)

Studio is designed to make AI development accessible whether you're working through a visual interface or programmatically via API.

Accessing Studio

Navigate to Studio in one of three ways:

1. Click the Studio button from your aiXplain dashboard
2. Go directly to studio.aixplain.com
3. Use the navigation menu in any aiXplain interface

When you first open Studio, you'll land in the Browse experience, where you can explore available AI assets.

Studio vs Console

Understanding the difference between Studio and Console helps you know where to go for specific tasks:

Studio	Console
Discover and explore AI assets	Manage account settings
Build and deploy agents	Manage teams and members
Test assets in real-time	View credits and billing
Integrate assets into workflows	Access analytics and usage reports
Connect integrations	Configure permissions and access

In short: Use Studio to work with AI assets. Use Console to manage your account and organization.

The Studio Experience

Browse

Browse is your gateway to aiXplain's asset marketplace. Here you can:

• Explore four types of assets: Models, Agents, Tools, and Integrations
• Filter and sort assets by category, vendor, popularity, price, and recency
• View detailed specifications for any asset
• Test assets using the Tryout feature
• Access integration code for programmatic use
• Save favorite assets for quick access

Each asset type serves a specific purpose:

• Models - AI models like LLMs, embeddings, speech-to-text, vision models, and more
• Agents - Autonomous AI agents that can use tools and models to accomplish tasks
• Tools - Capabilities that extend what agents can do (APIs, functions, operations)
• Integrations - Connectors that generate tools by linking external services

Learn more: Browse Documentation

Build

Build is where you create and deploy AI agents. You can build:

• Single agents - Agents that use tools and models to accomplish specific tasks
• Team agents - Multi-agent systems where specialized agents work together

The Build experience provides:

• Templates - Pre-built example agents to get started quickly
• Drafts - In-progress agents you're still configuring
• Agents - Your deployed agents (both single and team agents)

Building Single Agents

Single agents are ideal for focused tasks. They can:

• Use multiple tools and models
• Process different file types (audio, video, images, documents)
• Maintain conversation memory
• Be configured with custom instructions and behaviors
• Be integrated into your applications via API

Building Team Agents (Multi-Agent Systems)

Team agents orchestrate multiple specialized agents for complex workflows. They include:

• Sub-agents - Specialized agents that handle specific parts of a task
• Micro-agents - Built-in orchestration components:
  • Planner - Creates execution plans from user requests
  • Orchestrator - Routes work to the appropriate sub-agents
  • Inspector - Validates outputs against policies and criteria
  • Responder - Formats the final output

Team agents are ideal when you need:

• Complex multi-step workflows
• Specialized expertise for different parts of a task
• Governance and validation at multiple stages
• Coordination between different capabilities

Learn more: Build Documentation

Integrations

Integrations connect external services to aiXplain, generating tools you can use in your agents. aiXplain offers both standard integrations and premium integrations with specialized capabilities:

Premium Integrations

• SQLite - Connect databases for natural language querying
• Python Sandbox - Execute custom Python scripts securely
• MCP Server - Connect Model Context Protocol servers
• AIR Vector Database - Semantic search and document storage

Each integration generates a private tool that becomes available to your agents.

Learn more: Integrations Documentation

Key Concepts

Assets

Assets are the building blocks of AI development in aiXplain. Every model, agent, tool, and integration is an asset with:

• A unique Asset ID for programmatic access
• Public or Private visibility
• Specifications including description, pricing, and capabilities
• Integration code for use in applications

Learn more: Understanding Assets (in Concepts section)

Privacy

Assets can be Public or Private:

• Public assets - Visible to all aiXplain users, cannot be edited by you
• Private assets - Visible only to your team, fully editable and transferable

When you create agents, they start as private assets by default.

Learn more: Teams and Workspaces

Deployment

Every agent you create in Studio is automatically deployed as an asset. This means:

• It's immediately available via API
• You can chat with it directly
• Other team members can discover and use it
• It can be added to other team agents

Updating an agent redeploys it with your changes instantly.

Validation

Validation is how you test your agents. During validation, you can:

• See real-time execution of every component
• Track inputs and outputs at each step
• Monitor credit consumption
• View the agent's reasoning and decision-making
• Debug issues by clicking on specific components

Validation provides full transparency into how your agents work.

Getting Started

New to Studio?

1. Start with Browse - Explore available models, agents, and tools to understand what's possible
2. Try the Tryout feature - Test assets directly without writing code
3. Use a template - Build your first agent from a pre-configured example
4. Add simple tools - Start with one or two tools to understand agent behavior
5. Iterate - Use validation to refine your agent's instructions and configuration

Ready to Build?

1. Define your use case - What should your agent accomplish?
2. Choose single or team - Does this need one agent or multiple specialists?
3. Select your tools - What capabilities does your agent need?
4. Configure instructions - Describe the behavior you want
5. Test thoroughly - Use validation to ensure it works as expected

Going to Production?

1. Optimize performance - Reduce unnecessary tools and actions
2. Set up governance - Use Inspectors for team agents to enforce policies
3. Test edge cases - Validate behavior with unexpected inputs
4. Monitor usage - Track credit consumption and performance
5. Use the API - Integrate your agent into your application

Studio Interface Navigation

Top Navigation

• Studio logo (top left) - Return to Browse experience
• Browse / Build tabs - Switch between experiences
• Search bar - Find assets quickly
• Feedback button - Report issues or send feedback

Browse Navigation

• Asset type filters - Models, Agents, Tools, Integrations
• Sort options - Popularity, Price, Recency
• View toggle - Card view or List view
• Save/Favorites - Access your saved assets

Build Navigation

• Create new agent - Start building
• Templates / Drafts / Agents - Organized categories
• Agent Builder Canvas - Schema, Validation, Publish tabs
• Chat interface - Test your agents

Common Workflows

Discovering a Model for Your Application

1. Go to Browse
2. Filter by Models
3. Sort by Popularity or search for specific functionality
4. Click on a model card to view specifications
5. Test it in Tryout
6. Copy the integration code from API Integration tab

Building a Customer Service Agent

1. Go to Build → Create new agent
2. Configure name: "Customer Service Agent"
3. Write instructions describing how to handle inquiries
4. Add tools: Knowledge base search, Ticket creation, Email sender
5. Test with sample customer questions
6. Refine instructions based on validation results
7. Deploy and integrate via API

Creating a Research Team Agent

1. Go to Build → Create new agent
2. Configure as team agent
3. Add sub-agents:
   • Web Research Agent (with web search tools)
   • Data Analysis Agent (with data processing tools)
   • Report Writer Agent (with document creation tools)
4. Configure Inspector to check for citation requirements
5. Test with research questions
6. Monitor how Planner and Orchestrator coordinate the team

Connecting a Database

1. Go to Browse → Integrations
2. Find SQLite integration
3. Click Connect
4. Upload your database file
5. Name the generated tool
6. Add it to your agent in Build
7. Test natural language queries

Tips for Success

In Browse

• Save assets you use frequently for quick access
• Test assets thoroughly before integrating
• Check performance metrics for models (success rate, latency, uptime)
• Read descriptions carefully - they tell agents how to use assets

In Build

• Start simple and iterate - add complexity gradually
• Write clear, specific instructions for your agents
• Select only the tool actions you need (not "enable all")
• Stay within 10-15 tools per single agent for best performance
• Use validation extensively to understand agent behavior
• Name your agents and tools descriptively

For Team Agents

• Design sub-agents with focused, specific purposes
• Use Inspector strategically - not on every component
• Test with progressively complex queries
• Monitor validation traces to understand orchestration
• Be aware that editing sub-agents affects all teams using them

Support and Resources

Getting Help

• Feedback button - Available throughout Studio
• Report an issue - On any public asset card
• Documentation - Comprehensive guides for all features
• Support - Contact aiXplain support team

Learning Resources

• Templates - Example agents demonstrating different patterns
• Validation traces - Learn by observing execution
• Best practices guides - Optimizing performance and cost
• API documentation - Programmatic access details

What's Next?

Explore Browse

Learn how to discover and work with models, agents, tools, and integrations.

Browse Documentation

Start Building

Create your first single agent or multi-agent system.

Build Documentation

Connect Integrations

Extend your agents with premium integrations.

Integrations Documentation

Understand Core Concepts

Deep dive into assets, teams, and credits.

Concepts Documentation (coming soon)

---

Ready to get started? Head to studio.aixplain.com and begin exploring!