Skip to content

Introduction ​

Gnosari is an AI-powered data collection platform that enables businesses to deploy conversational AI agents that collect leads, feedback, orders, and other structured data through natural conversations. Agents can be published on custom joina.chat URLs or embedded on any website via the chat widget.


What is Gnosari? ​

Gnosari allows you to create and deploy AI agents that collect data automatically through conversation. Instead of traditional forms, users interact naturally with an AI agent that asks questions, understands context, and extracts structured information in real-time.

Core value proposition: Turn every conversation into actionable data without forcing users through rigid forms.


Platform Architecture ​

Users create and configure agents in the Gnosari UI, publish them to public chat interfaces (joina.chat or embedded widget), and the Gnosari Engine handles AI orchestration and automatic data extraction into a PostgreSQL database.


Key Concepts ​

Agents ​

AI assistants that handle conversations and collect data. Each agent has:

  • Purpose: Pre-configured type (Customer Support, Lead Collector, Booking Assistant, etc.)
  • Instructions: Behavioral guidelines and conversation goals
  • Knowledge: Optional RAG knowledge from websites, documents, or databases
  • Capabilities: Advanced reasoning, image understanding, task delegation, memory
  • Tools: Optional external tool integrations
  • Traits: Reusable personality and communication patterns

Teams ​

Multi-agent collaboration systems where multiple AI agents work together on complex workflows. Teams have:

  • Agent Roles: Manager (coordinates), Orchestrator (routes), Member (executes)
  • Connections: AI Delegation (manager → member with feedback loop), Handoff (one-way transfer)
  • Visual Designer: Flow-based canvas for designing team structures

Knowledge Sources ​

RAG (Retrieval-Augmented Generation) system for giving agents access to external information:

  • Sources: Websites, PDFs, DOCX files, Markdown, APIs, databases, cloud storage
  • Semantic Search: AI-powered search with relevance scoring

Sessions ​

Chat conversation history between users and AI agents/teams:

  • Messages: User, Assistant, Function Call, Function Result, Reasoning
  • AI Insights: Automatic summary, sentiment analysis, key topics, action items
  • Real-time: WebSocket-based streaming responses

Traits ​

Reusable behavioral patterns and personality characteristics:

  • Categories: Communication, Technical, Customer Service, Management, Creative, Analytical
  • Weight System: Low (0-1.0), Normal (1.0-1.1), High Impact (1.1+)
  • Composable: Multiple traits can be combined on a single agent

Chat Themes ​

Visual and behavioral configurations for the public chat interface:

  • Welcome Experience: Welcome modal, agent greeting, empty state
  • Interface: Header, chat banner, feature sidebar
  • Conversation Prompts: Starter prompts, auto messages, follow-ups
  • Color Theme: 22 brand color palettes
  • Background Pattern: WhatsApp-style decorative patterns

Automations ​

Event-driven triggers that automatically run AI agents when specific events occur:

  • Triggers: Event pattern matching (exact, prefix, wildcard)
  • Actions: Run specified agent or team
  • Message Templates: Jinja2 templates with event payload variables
  • Priority Levels: Critical, High, Normal, Low

Feature Summary ​

FeatureDescriptionDocumentation
AgentsCreate and manage AI agents with 12+ pre-configured purposesManaging Agents
Build Wizard4-step wizard for creating agents via conversation with AIBuild Wizard
Agent ConfigurationAdvanced form for full agent customizationAgent Configuration
TeamsMulti-agent collaboration with visual designerManaging Teams
KnowledgeRAG system with documents, websites, APIs, databasesSources
SessionsChat history with AI-generated insightsSession Management
Chat SystemReal-time WebSocket chat with streaming responsesChat System
Chat ThemesVisual and behavioral customization for public chatChat Themes
TraitsReusable personality and behavioral patternsManaging Traits
AutomationsEvent-driven agent triggersAutomation Configuration
WidgetEmbeddable chat widget for websitesWidget Embedding
Scheduled TasksTask scheduling with cron expressionsScheduled Tasks
SettingsAccount settings, subscription, API keysSettings

Who Uses Gnosari? ​

Business Use Cases ​

  • Sales Teams: Lead qualification, appointment scheduling, product information
  • Customer Support: 24/7 FAQ handling, ticket creation, issue resolution
  • Marketing: Event registration, feedback collection, survey responses
  • HR/Recruiting: Candidate screening, interview scheduling, job information
  • E-commerce: Order collection, product recommendations, customer service

Personal Use Cases ​

  • Content Creators: Fan engagement, content ideas, social media management
  • Educators: Study assistance, tutoring, assignment help
  • Personal Branding: AI clone for 24/7 availability, social media engagement

Getting Started ​

New to Gnosari? Start with the Getting Started Guide to create your first AI agent.

For developers working on the UI codebase, see the Development section.


Platform Capabilities ​

AI Capabilities ​

  • Advanced Reasoning: Step-by-step logical reasoning for complex tasks
  • Image Understanding: Analyze and understand images in conversations
  • Task Delegation: Agents can delegate to other agents in teams
  • Conversation Memory: Remember context from previous conversations (Beta)

Data Collection ​

  • Structured Data Extraction: Automatically extract typed data from conversations
  • Entity Types: Define custom data schemas for collection
  • Real-time Analytics: Track conversation metrics and data collection progress
  • Export: Export collected data to CSV, JSON, or via API

Integration Options ​

  • Public Links: joina.chat/your-agent-handle
  • Website Widget: Embeddable chat widget (bubble, sidebar, drawer modes)
  • API Access: RESTful API for programmatic agent management
  • Webhooks: Event-driven automation triggers

Next Steps ​

  1. Getting Started: Create your first agent in under 5 minutes
  2. Features: Explore all platform features in depth
  3. Development: Technical documentation for developers
  4. Troubleshooting: Common issues and solutions