Multi-tenant SaaS
Isolated tenant ownership and authorization.
Retrieval-augmented generation, operationalized
Turn trusted knowledge into grounded AI answers.
Create tenant-isolated chatbots, process valid UTF-8 .txt knowledge, retrieve relevant context through Qdrant, and answer with OpenAI or Gemini.
Knowledge AssistantFrontend-only product demo
Static demonstration — no API or usage call
Multi-tenantGrounded RAGOpenAI + GeminiQdrant searchAPI + WidgetMonitoring
One connected platform
From knowledge file to useful answer.
Manage chatbots and valid UTF-8 .txt documents, generate embeddings, perform tenant-filtered Qdrant retrieval, and deliver grounded answers without stitching together separate application layers.
Upload knowledge
Process embeddings
Retrieve context
Generate grounded answers
Deploy API or widget
Review conversations & analytics
Capabilities
The operational pieces a RAG SaaS needs.
Chatbot Management
Create, configure, and control chatbot availability.
Document Processing
Private valid UTF-8 .txt ingestion with queue states.
OpenAI + Gemini Embeddings
Provider-registry based batch embeddings.
Qdrant Vector Search
Filtered retrieval by tenant, chatbot, and ready document.
Grounded RAG Answers
Answers constrained to retrieved knowledge.
Public Chat API
Rate-limited public answering with safe responses.
Embeddable Widget
Framework-free Shadow DOM website widget.
Conversation History
Idempotent visitor-grouped conversation storage.
Analytics & Reports
UTC trends, usage, document, and revenue reporting.
Notifications
Queued database and email notifications.
API Keys & Webhooks
Scoped keys and signed outbound event delivery.
Security & Monitoring
Headers, CSP, health checks, metrics, and alerts.
Usage Limits
Transactional plan feature enforcement.
Subscription Billing
Plan pricing and Razorpay checkout.
Light / Dark Theme
Accessible appearance across the landing experience.
How it works
A traceable path from source to response.
- Step 1
Upload
Add valid UTF-8 .txt knowledge through private storage.
- Step 2
Chunking
Split source text into bounded retrieval chunks.
- Step 3
Embedding
Generate vectors through OpenAI or Gemini.
- Step 4
Qdrant Search
Find tenant-filtered relevant context.
- Step 5
LLM
Generate from the selected knowledge only.
- Step 6
Grounded Answer
Return a concise answer with citations.
Interactive product demonstration
See grounded retrieval in motion.
This frontend-only demonstration uses static sample content. It does not call the Public Chat API, OpenAI, Gemini, Qdrant, or consume usage.
User asks: How can I reset my account password? Assistant answers according to the uploaded knowledge base and cites the account guide.
UserHow can I reset my account password?
Source found
[1] Account guide
Password recovery instructions found in the uploaded knowledge.
Assistant
According to the uploaded knowledge base, use the password reset link on the tenant login page. [1]
Pricing
Plans configured for the way you operate.
Prices, currencies, billing cycles, discounts, and limits come directly from active platform plans.
Premium
For growing businesses
- Chatbots10
- Documents200
- Monthly questions25,000
- Storage (MB)20,480
- Team members10
Standard
For small teams
- Chatbots1
- Documents20
- Monthly questions2,000
- Storage (MB)2,048
- Team members2
Plan comparison
Compare configured feature limits.
| Feature | Premium | Standard |
|---|---|---|
| Chatbots | 10 | 1 |
| Documents | 200 | 20 |
| Monthly questions | 25,000 | 2,000 |
| Storage (MB) | 20,480 | 2,048 |
| Team members | 10 | 2 |
Analytics UI preview
Operational signals without opening private content.
This static preview uses synthetic values—not tenant records—to represent conversation, question, fallback, chatbot, document, revenue-by-currency, and provider/model reporting.
Conversations—
Questions—
Fallback rate—
Document status · chatbot usage · revenue by currency · provider/model usage
Integrations
Connected to the tools already in the platform.
OpenAIEmbedding and answer providers
GeminiEmbedding and answer providers
QdrantFiltered vector retrieval
RazorpaySubscription checkout
Laravel APIVersioned tenant integrations
JavaScript WidgetFramework-free website chat
WebhooksSigned outbound events
AnalyticsUsage and operational reports
Built with trusted technologies
OpenAIGeminiQdrantLaravelRazorpayREST API
Integration names describe platform compatibility and do not imply endorsement.Security by implementation
Practical protections throughout the request path.
These are implemented platform controls, not claims of formal certification or legal compliance.
- Tenant isolation
- Encrypted secrets
- Scoped API keys
- HMAC-signed webhooks
- Webhook SSRF protection
- Layered rate limiting
- CSP and security headers
- Correlation IDs
- Health checks and monitoring
- Privacy-safe Super Admin monitoring
FAQ
Clear answers about the current platform.
What is RAG?
Retrieval-augmented generation finds relevant knowledge first, then supplies that context to an AI model so its answer can remain grounded in the retrieved sources.
Which AI providers are supported?
OpenAI and Gemini are supported for embeddings and answer generation through provider registries.
Which document formats are supported?
Valid UTF-8 .txt files only in the current version. PDF, DOCX, and OCR ingestion are not advertised or provided here.
Can the chatbot be embedded?
Yes. The platform provides a framework-free JavaScript widget and a Public Chat API.
Is tenant data isolated?
Yes. Tenant-facing queries, API keys, chatbots, documents, conversations, and retrieval filters use tenant-scoped ownership.
How are usage limits handled?
Plans can enable, disable, or numerically limit features. Unlimited, zero, and positive allowances are enforced using the configured feature semantics.
Is an API available?
Yes. A public answering endpoint and scoped versioned tenant API endpoints are implemented.
Are outbound webhooks available?
Yes. Tenant owners can configure allowlisted HTTPS webhook events with HMAC signatures, queued delivery, retries, and SSRF checks.
How does pricing work?
Active plan prices are configured by billing cycle and currency. Checkout remains subject to tenant-owner authorization and the existing Razorpay flow.
Build on trusted context