Everything you need to ship
Production-grade AI assistants, from ingestion to analytics.
Multi-format ingestion with RAG-grounded answers
Upload PDFs, DOCX files, or scrape URLs. We automatically parse, chunk, and embed your content into a private vector database. Retrieval-Augmented Generation ensures your bot only answers from your data. Zero hallucinations, citations included.
- ✓pgvector for blazing-fast semantic search
- ✓Automatic chunking and embedding with text-embedding-004
- ✓Live context injection via API for dynamic data
Real-time streaming with smart model routing
Token-by-token streaming with the Vercel AI SDK. Simple questions use fast, cheap models. Complex queries escalate to premium models automatically. Configure icebreaker questions to guide conversations toward high-value topics.
- ✓Sub-200ms first token latency
- ✓Gemini, Groq, DeepSeek, OpenAI auto-routing
- ✓Customizable suggested questions
Full white-label with embeddable widget
Per-bot names, icons, accent colors, and tone. Match your brand perfectly. Drop a single script tag on any website: floating trigger button, configurable positioning, iframe isolation for security.
- ✓Custom branding per bot
- ✓One-line embed code
- ✓TypeScript SDK for programmatic control
Multi-agent collaborative document editing
Deploy specialized AI agents to help write, review, and refine professional documents. Every suggestion is grounded in your knowledge sources and presented as inline edits you can accept or reject, like Google Docs suggestions but powered by your proprietary data.
- ✓Deploy multiple AI agents per document
- ✓Inline suggestions, sidebar chat, and document-wide review
- ✓Ghost text auto-suggestions as you type
- ✓Real-time presence, version history, export to PDF/Word/Markdown
Analytics and API key management
Track message volume, top queries, satisfaction scores, and conversion rates. Generate, rotate, and revoke API keys with per-key usage tracking. Enterprise plans can bring their own LLM API keys for full cost control.
- ✓Real-time analytics dashboard
- ✓Scoped API keys with rate limiting
- ✓Webhook events for automation
Multi-platform commerce integration
Connect your WooCommerce, Shopify, PrestaShop, or Magento store with secure OAuth. Your AI assistant can search products, answer inventory questions, manage shopping carts, and track orders,all directly in chat. WordPress plugin available for one-click installation.
- ✓WooCommerce, Shopify, PrestaShop, Magento support
- ✓Product search, cart management, order tracking in chat
- ✓Secure OAuth pairing with tool policy engine
- ✓WordPress plugin for one-click deployment
Staff console with human handoff
When AI can't handle a query, live staff can take over seamlessly. Monitor active conversations, view full context, use suggested replies and quick templates, and maintain a complete audit trail for compliance. Every action is logged: user messages, bot responses, staff takeovers, and tool executions.
- ✓Live agent takeover with session state management
- ✓Suggested replies and quick reply templates
- ✓Full conversation audit trail for compliance
- ✓Real-time conversation monitoring dashboard
75+ widget styling options with live preview
Customize every visual aspect of your chat widget: colors, typography, layout, animations, branding, and floating button configuration. Choose from 3 presets (Glass, Solid, Minimal) or design your own. Live preview editor shows changes instantly before deployment.
- ✓75+ styling options: colors, fonts, layout, animations
- ✓3 visual presets: Glass, Solid, Minimal
- ✓Live preview editor with instant updates
- ✓Per-bot styling isolation for multi-brand deployments
Enterprise-grade platform hardening
Mission-critical reliability with atomic billing, XSS protection, rate limiting, and zero-downtime recovery. Built to handle enterprise-scale deployments with 99.9% uptime and comprehensive error handling.
- ✓Atomic message counting prevents billing overages
- ✓HTML sanitization on all AI responses (XSS protection)
- ✓Per-org rate limiting and resource creation limits
- ✓Billing cycle recovery with automatic error boundaries
Dynamic page context for grounded answers
The embeddable widget automatically captures the host page's URL, title, product data (name, price, SKU), and content sections from standard HTML markup. This context is threaded into the AI's system prompt so every answer is grounded in what the visitor is currently viewing. Product-aware icebreakers appear when e-commerce data is detected, in 7 languages. Zero configuration, works instantly on any page.
- ✓Auto-detects product data: name, price, SKU from DOM
- ✓Extracts page headings and content sections for full understanding
- ✓Localized icebreakers: 'I have a question about this product' in 7 languages
- ✓Privacy-safe: context stays in session, graceful degradation on all pages