Documentation

An illustrated user manual: step-by-step instructions, real interface screenshots with sample data, and checks for each feature.

New to AI Studio?

Build your first application

Follow the steps from creating an application to testing and publishing it.

Getting started 4 guides

Create your first application

Build a chatbot, agent, chatflow or workflow.

Explore model providers

Compare the provider catalog before configuring models.

Connect AI models

Choose the models your workspace can use.

Manage your account

Keep sign-in details and personal preferences up to date.

Application design 5 guides

Configure applications and multimodal features

Configure inputs, conversation features, speech, image generation and business processes.

Design agent business processes

Separate goals and instructions into processes selected for each user request.

Variables and input forms

Collect structured data before an application runs.

Speech and image generation

Enable speech input, spoken answers and image generation with supported models.

Version history and restoration

Restore earlier configuration into a draft while controlling publication.

Knowledge & data 8 guides

Build a knowledge base

Ground answers in your own documents.

Manage documents and chunks

Track ingestion, inspect original documents and manage chunks used for retrieval.

Folders, filtering and moving documents

Organise knowledge documents and act on the intended scope.

Inspect and edit knowledge chunks

Edit the content actually used for retrieval.

Test and tune retrieval

Inspect retrieved context before using it in answers.

Design knowledge ingestion pipelines

Choose a template and customise how data enters a knowledge base.

Connect data sources

Reuse source connections for document ingestion.

Connect external knowledge

Connect an existing retrieval service in your organisation.

Workflows & automation 8 guides

Design visual workflows

Connect processing steps and inspect each run.

Workflow node reference

Choose and configure the 28 node types available in the editor.

Run, trace and compare workflows

Locate failing steps through each node’s actual inputs and outputs.

Import/export DSL and environment variables

Move application structure between environments and configure secrets at the destination.

Reuse snippets and edit the canvas

Save reusable structures and use the canvas editing tools.

Automate with triggers

Start workflows from events or schedules.

Webhooks, schedules and event retries

Attach triggers to workflows and track incoming events.

Review approval requests

Keep a human decision in approval-gated flows.

Tools & integrations 5 guides

Give agents tools

Connect built-in tools, HTTP APIs and MCP servers.

Configure API tools

Let agents call APIs registered in the workspace.

Connect MCP servers

Discover and select MCP operations for agents.

Use workflows as tools

Reuse a workflow through a tool interface.

Use extension APIs for moderation

Publishing & quality 13 guides

Publish and share

Deliver a tested version through the right channel.

Web apps, embedding and chat experience

Deliver a published application through a link or embedded widget.

Connect messaging channels

Connect a published application to a supported communication channel.

Use the application API and MCP

Integrate published apps into another system.

Workspace API keys and MCP capabilities

Issue scoped keys for integrating systems.

Read logs and execution traces

Trace application results back to conversations or execution steps.

Set up canned answers

Provide consistent answers to explicitly matched questions.

Evaluate and compare quality

Use test datasets to track application quality across runs.

Export training data

Prepare application data before submitting a training job.

Monitor usage and quality

Understand runs, costs, feedback and audit events.

Manage long-term memory

Review and govern information retained across interactions.

Review memories and their evidence

Decide what may be reused in later AI turns.

Configure, retain and disable memory

Control extraction frequency, retention and data when disabling memory.

Workspace management 10 guides

Manage your workspace

Organize members, roles, API keys and quotas.

Invite members and control access

Place people in the correct workspace, role and departments.

Create roles and assign permissions

Define workspace permissions around team responsibilities.

Token and request quotas

Set usage limits for the appropriate scope and period.

Configure data protection

Align AI usage with your organization’s data policies.

Manage outbound connections

Configure allowed destinations for tools and external services.

Set up branding, storage and identity

Configure workspace services from dedicated settings pages.

Manage S3 storage

Monitor the bucket, storage use and unreferenced objects.

Configure single sign-on

Connect an identity provider and map user groups.

Fine-tune a model

Submit and track training through configured providers.