Agent Basics

Learn how to interact with AI agents, select models, and perform basic operations.

AI agents are the central workspace of Jitera. Simply send a message to have agents handle code analysis, document creation, web research, file generation, and more.

Getting Started

1. Open a Chat

   Click New Chat in the sidebar, or click an agent name to start a conversation.

2. Select an Agent

   Choose an agent from the Agents section in the sidebar. Each agent has different strengths. See Choosing the Right Agent for details. You can also create your own agents — see Managing Agents.

3. Select an AI Model

   Click the model name displayed below the agent name to open the model selection dialog. Each model shows its provider and data processing region.

   

4. Send a Message

   Type your message in the input field and press Enter to send. Use Shift + Enter for line breaks.

5. Return to Past Threads

   Open previous conversations from Threads in the sidebar.

@Mentions

Type @ in the chat input to reference resources within your project.

Type	Purpose
Documents	Share project documents with the agent
Members	Mention or invite team members in group chat
Agents	Invoke a different agent in group chat

/Commands (Skills)

Type / in the chat input to see available skill commands. Skills extend the agent’s specialized capabilities — selecting one sends your message with that skill activated.

Common commands:

Command	Description
/deep-research	Perform in-depth research and analysis across multiple sources
/excel-generator	Create and edit Excel (.xlsx), CSV, and TSV files

• Type characters after / to filter commands
• Use arrow keys to select and Enter to confirm
• Skills not enabled for the agent are shown with an “OFF” badge

For more on skills, see Skills.

AI Model Selection

Available models vary by organization settings. The main providers are OpenAI, Anthropic, and Google.

Each model displays its context window size (how much text it can process) and data processing region. Your organization’s administrator configures which models are available.

What Agents Can Do

Agents automatically choose the right tools based on your message.

Capability	Description
Search and read documents & code	Search across project files to understand content
Create and edit documents	Create new documents or update existing ones
Web search and page reading	Research the latest information online and verify sources
Python execution	Data processing, analysis, and script execution
File generation	Create downloadable Excel (.xlsx), CSV, and other files
Context search	Reference learned information from past conversations
Skill activation	Trigger specialized capabilities like Excel generation or deep research as needed
MCP integration tools	Use tools from connected external services

For details, see Tool Calls / Actions.

Workspace

Agents can access three areas within a project.

Area	Content	Access
Documents	Project documents	Read & write
Source code	Code from connected Git repositories	Read only
Uploads	Files attached to chat	Read only

For connecting source code, see Code Sources.

Related

• Choosing the Right Agent — Which agent to use and when
• Chat Threads — Managing and organizing conversations
• File Upload in Chat — Attaching files to messages
• Tool Calls / Actions — Understanding agent tool execution