Getting Started with Velora!
Build and deploy your first AI agent in under 5 minutes.
1. Create an account
Sign up at velora.app/signup. No credit card required.
2. Create an agent
From the dashboard, click "+ New Agent" or pick a template. Give your agent a name, type, and personality.
3. Train with documents
Go to Knowledge Base and upload PDFs, DOCX, TXT, CSV files or add URLs. Your agent learns from them automatically using hybrid vector search (semantic + keyword).
4. Test in the playground
Open the Playground tab and chat with your agent. Test questions, check sources, and refine the personality.
5. Deploy
Go to Deploy, generate an API key, and choose your deployment method: website widget, mobile PWA, Telegram bot, or API.
API Reference
Authenticate with an API key using the X-API-Key header.
/api/widget/{agent_id}/chatSend a message and receive a streaming response (SSE).
Request
curl -X POST "https://your-backend.com/api/widget/{agent_id}/chat" \
-H "Content-Type: application/json" \
-H "X-API-Key: vk_your_api_key" \
-d '{
"message": "What is your return policy?",
"conversation_id": null
}'Response (SSE stream)
data: {"type": "token", "content": "Our"}
data: {"type": "token", "content": " return"}
data: {"type": "token", "content": " policy"}
data: {"type": "metadata", "conversation_id": "uuid", "sources": ["faq.pdf"], "response_time_ms": 1234}
data: [DONE]/api/widget/{agent_id}/configGet agent configuration (name, greeting, color). Requires API key.
curl "https://your-backend.com/api/widget/{agent_id}/config" \
-H "X-API-Key: vk_your_api_key"Response
{
"agent_name": "Support Bot",
"greeting_message": "Hi! How can I help?",
"widget_color": "#6366f1",
"widget_position": "bottom-right",
"enable_voice": false,
"enable_vision": false
}/api/widget/{agent_id}/public-configGet minimal public info (no API key required). Used by the PWA landing page.
curl "https://your-backend.com/api/widget/{agent_id}/public-config"Website Widget
Embed a chat bubble on any website with a single script tag.
Installation
Add this script before the closing </body> tag:
<script src="https://your-backend.com/widget/widget.js" data-agent-id="your-agent-id" data-api-key="vk_your_api_key" data-host="https://your-backend.com" ></script>
Customization
The widget automatically uses your agent's configured color and position. Change these in Settings → Widget tab in the dashboard.
Supported options
widget_color: Hex color for the chat bubble and header
widget_position — bottom-right or bottom-left
Mobile App (PWA)
Share your agent as a standalone mobile app. No app store needed.
How it works
Each agent gets a unique shareable URL:
https://your-app.com/a/{agent_id}?key=vk_your_api_keyWhen users open this on mobile:
1. They see a full-screen chat interface with your agent's branding
2. They can tap "Add to Home Screen" in the browser menu
3. The agent installs as a standalone app with its own icon and name
4. Each agent installs as a separate app — completely isolated
Note on HTTPS
The automatic install prompt only appears on HTTPS. On localhost, users need to manually use "Add to Home Screen" from the browser menu.
Telegram Bot
Deploy your agent as a Telegram bot.
1. Create a bot
Open Telegram, search for @BotFather, send /newbot, and follow the instructions.
2. Connect to Velora!
Go to your agent's Deploy page, scroll to Telegram Bot, paste the token from BotFather, and click Connect.
3. Start the bot
Click Start on the dashboard. Your bot is now live and uses the same RAG pipeline, tools, and personality as your agent.
Commands
/start: Start the conversation
/help: Show help message
/new: Start a new conversation
Integrations
Connect external services so your agent can read and act on real data.
Gmail
Read, search, draft, and send emails
Read (auto), Send (approval required)
Google Calendar
Read events, find free time, create events
Read (auto), Create (approval required)
Google Drive
Search and list files
Read-only (auto)
YouTube
Search videos, get video details
Read-only (auto)
Slack
List channels, read history, send messages
Read (auto), Send (approval required)
HubSpot
Search contacts and deals, create records
Read (auto), Create (approval required)
All integrations use OAuth 2.0. Connect from your agent's Integrations tab. Tokens are encrypted at rest using Fernet symmetric encryption.
Agent Tools
Every agent has access to built-in tools. The AI decides when to use them.
web_searchSearch the web using DuckDuckGo
AUTOcalculatorEvaluate math expressions
AUTOget_datetimeGet current date, time, and timezone
AUTOread_emailsRead recent emails from Gmail
AUTOsend_emailSend an email via Gmail
APPROVALread_eventsRead calendar events
AUTOcreate_eventCreate a calendar event
APPROVALsearch_driveSearch Google Drive files
AUTOsearch_youtubeSearch YouTube videos
AUTOsend_slack_messageSend a Slack message
APPROVALcreate_hubspot_contactCreate a HubSpot contact
APPROVALPermission levels
AUTO: Agent uses the tool automatically without asking.
APPROVAL: Agent asks the user for permission before executing.
Custom APIs
Connect any REST API as an agent tool - your CRM, internal services, or third-party APIs.
How to add a custom API
Go to your agent's Integrations page, scroll to Custom APIs, and click "+ Add API".
Configuration fields
Tool Name: Unique identifier (e.g., get_weather). No spaces.
Description: What the API does. The AI reads this to decide when to use it.
Method: GET, POST, PUT, or DELETE.
URL: Full endpoint URL. Use {param} placeholders for dynamic values.
Parameters: JSON array defining what inputs the AI should provide.
Headers: JSON object for auth headers, API keys, etc.
Body Template: JSON template for POST/PUT requests.
Response Path: JSONPath to extract useful data from the response.
Example: Weather API
{
"name": "get_weather",
"description": "Get current weather for a city",
"method": "GET",
"url": "https://api.weather.com/v1/current?city={city}",
"parameters": [
{
"name": "city",
"type": "string",
"description": "City name",
"required": true
}
],
"headers": {
"X-Api-Key": "your-weather-api-key"
}
}