Skip to content

Quick Start

Get started with ELITEA by sending your first chat message. This hands-on guide will walk you through the essentials of using Chat, the heart of the ELITEA platform.

Step 1: Open Chat

  1. Navigate to ELITEA: Open your browser and go to https://elitea.ai
  2. Sign In: Enter your account credentials.
  3. Access Chat: Click Chat in the left sidebar (or you'll be redirected there automatically).

Chat Interface Overview

You'll see the Chat interface with three main areas:

  • Left sidebar: Your conversations list
  • Center area: Message history and input box
  • Right sidebar: Participants panel

Step 2: Send Your First Message

  1. Locate the message input box at the bottom of the center area.
  2. Select an LLM Model: Look for the LLM model selector at the bottom of the chat interface
  3. Type Hello or What model are you and what are your capabilities?
  4. Click the Send button (paper airplane icon) or press Enter.

First Message

Success Indicator

If you see a response from the AI, you've successfully sent your first message!

Step 3: Connect a Toolkit (Jira)

Connecting tools enhances your chat capabilities by giving the AI access to external services like Jira, GitHub, or Confluence.

Before you begin

You'll need:

  • Jira account credentials (base URL, username, API token)

How to connect Jira:

  1. Create a Toolkit from Canvas:

    • In the PARTICIPANTS panel on the right, find toolkits
    • Click the + Add toolkit icon
    • Click + Create new Toolkit
    • Select Jira as the toolkit type
    • Enter a name (e.g., JiraAssistant)

  2. Create a Credential (when configuring the toolkit):

    • In the CONFIGURATION section, click the credentials dropdown
    • Select + New credentials (or + New private credentials)
    • You'll be redirected to the Create Credentials page
    • Select Jira as the credential type
    • Give it a name like My Jira Credentials
    • Fill in your Jira Base URL, and Token(Auth: Bearer) or Api Key and Username (Auth: Basic)
    • Click Save

  3. Return to Toolkit Creation:

    • Go back to the toolkit creation canvas
    • Click the refresh icon next to the Jira configuration field
    • Your newly created credential will appear in the list
    • Select it from the dropdown

  4. Complete Toolkit Configuration:

    • Select an Embedding Model (e.g., text-embedding-3-small)
    • Choose a Vector Store (PostgreSQL with pgvector extension)
    • Enable the tools you need (e.g., Search using JQL, Create issue)
    • Click Create
    • Click X to close the Toolkit Canvas interface
    • The toolkit will appear in your PARTICIPANTS panel

  5. Use it:

    • Type: Find all open bugs in project ABC
    • Click Send
    • The toolkit will search Jira and return results

Connect Toolkit

Step 4: Send an Attachment

Attach files or images to your chat messages for AI-powered analysis.

Supported File Types

  • Images: JPEG, JPG, PNG, GIF, and more
  • Documents & Data: JSON, TXT, MD, YAML, and more
  • Maximum: 10 images per message
  • Size limits: Vary by AI model (e.g., 5MB for Anthropic, 20MB for OpenAI)

How to Attach Files

  1. Enable Attachments (first time only):

    • Click the paperclip icon in the message input area
    • The Attachment settings popup appears
    • Select an existing Artifact Toolkit or create a new one
    • Click Save

  2. Attach Files (three ways):

    • Click: Click the paperclip icon → select files
    • Drag & Drop: Drag files from your computer into the message input area
    • Paste: Copy an image (Ctrl+V or Cmd+V) and paste it

  3. Add Text Prompt (required):

    • Type a message describing what you want the AI to do
    • Example: Analyze this diagram and explain the workflow

  4. Send:

    • Click the Send button
    • The AI will analyze your attachment and respond

Attach Files

Step 5: Create an Agent from Chat (Optional)

Once you've had a successful chat session, you might want to reuse the same setup. Creating an agent saves your configuration for future use.

How to create an agent

  1. Have a conversation with the desired tools and settings.
  2. Create Agent from Canvas:
    • In the PARTICIPANTS panel on the right, find Agents
    • Click + Add agent icon
    • Click Create new agent
    • This opens the Agent Canvas interface
  3. Configure Your Agent:
    • Give your agent a name (e.g., Jira Assistant)
    • Add a description of what it does (e.g., Helps manage Jira tickets and queries)
    • Add instructions that define its behavior (e.g., You are a Jira assistant. Help users search for tickets, create issues, and analyze project data. Always use JQL syntax for searches.)
    • Click Save to proceed to advanced configuration
  4. Add Toolkits
    • In the advanced configuration, select toolkits to attach to your agent
    • Configure LLM model and settings
    • Click Save
    • Click X to close the Agent Canvas interface
  5. Use Your Agent:
    • Select your agent in the PARTICIPANTS section (it will be highlighted)
    • Type a message and send (e.g., Find all high-priority tickets updated this week)

Create Agent

Benefits of Agent Creation

  • Reusability: Use the same setup across multiple chats
  • Consistency: Same instructions and tools every time
  • Sharing: Share with teammates in team projects
  • Efficiency: Skip setup and start working immediately

Step 6: Use Public Agents (Optional)

Explore pre-built agents created by the ELITEA community to discover new capabilities.

How to Find and Use Public Agents

  1. In the PARTICIPANTS panel on the right, find agents
  2. Click the + icon
  3. Browse the list of available agents
  4. Look for the "Public" badge to identify community-shared agents
  5. Select a public agent to add it to your conversation
  6. Click the agent in the PARTICIPANTS list to activate it
  7. Type your message or use conversation starters if available
  8. The agent responds based on its specialized instructions and capabilities

Public agent

Learning from Public Agents

Public agents are a great way to learn how others are using ELITEA. You can:

  • View their configurations (read-only) by clicking the ⚙️ settings icon
  • Understand their prompt engineering and structure
  • Create your own versions with customizations
  • Combine concepts from multiple public agents

Useful Links