§ Platform · Features

Every feature,
in one place.

Everything behind Qlar's solutions: how agents talk and remember, what they know, how they connect and act, and how you keep them governed.

Conversation & Personalization

Context-aware conversations
Natural, Personalized Chats
Ensure conversations flow naturally, in the user's language, from the very first message.
  • Define the assistant's personality, interaction style, and language (fixed, or detected from the user)
  • Greet each visitor personally with welcome-message variables and conditions (name, time of day, day of the week)
  • Offer example questions before the first message and clickable follow-up suggestions after each answer
  • Let users tap quick-reply widgets or speak with built-in voice dictation instead of typing
  • Let a host website send its own opening line when the chat starts
Personalization & memory
New
Remembers Every Person
The agent remembers what you decide it may remember about each person, and nothing else.
  • Declare memory fields, such as preferred branch, language, or allergies, and the agent fills them from conversation
  • Nothing is captured outside the fields you define; special-category data needs an Owner's approval
  • Every value keeps the quote it came from, when it was captured, and when it expires
  • Let people correct what is remembered, and decide which staff may see personal data; every access is logged
Setup in minutes
Brand-Ready Launch
Launch a chat assistant that matches your brand in minutes.
  • Set or generate assistant name, avatar, and logo
  • Customize colors and fonts to match your corporate identity or choose from pre-built templates
  • Fine-tune every UI color token individually or drop in custom CSS for full white-label control
  • Publish a public About page with a rich-text description, social media links, and tap-to-contact details (email, phone, WhatsApp) so visitors can learn more before they chat

Voice & Visual Experience

Voice-to-voice conversations
Turn Any Chat Into a Live Voice Call
Let users talk to your agent naturally, with the same knowledge, tools, and rules as text chat.
  • Immersive sound-wave view or captioned transcript; users can switch anytime, mid-call
  • Two conversation styles: turn-by-turn, or natural conversation (beta) for fluid back-and-forth
  • Choose a voice persona and a speaking style (Everyday, Expressive, or Formal), and tune voice detection or push-to-talk
  • Users can attach photos and documents during a call
  • Embed a one-tap "Start Voice Call" button right in your welcome message
Charts, tables, maps & more
Show, Not Just Tell
Let your agent present information visually, in text chat and voice calls alike.
  • Charts, tables, maps, formulas, and cited source pages appear on a dedicated visual canvas
  • Force specific content types onto the canvas instead of having them read aloud
  • On WhatsApp, the assistant can send the current canvas as a message
  • Works the same way whether the user is typing or on a live voice call

Knowledge & Skills

Accurate, trusted answers
Manage Agent's Knowledge
Supply your agent with your specialized information to ensure it delivers accurate, up-to-date answers and fosters trust.
  • Upload documents, paste website URLs, or link videos so the agent draws from your curated knowledge base
  • Restrict responses to your knowledge base so you can keep answers concise and accurate
  • Display citations inline or as footnotes, and link to the original source to let users verify information easily
  • Extract knowledge from text, tables, and images to acquire more detailed data and visual insights
  • Pick from multiple ingestion strategies (semantic, skimming, regulatory, fact-sheet, or adaptive), plus an AI reranking stage to sharpen retrieval accuracy
Procedures the agent follows
New
Skills & Artifacts
Teach the agent how to handle specific situations (a refund, an intake, an inspection), and give it the files it needs.
  • Write a skill as a procedure for one kind of situation, with a plain description of when to use it
  • The agent opens a skill only when the conversation calls for it, by keyword or by its own judgment, and lets it go when the topic moves on
  • Give skills folders of reference notes, templates, or pictures to read or show
  • Attach skills to automations so scheduled runs follow the same procedure
Always up to date
Web Search & Live Data
Give your agent access to live web data so it can answer questions that go beyond your static knowledge base.
  • Enable web search to let the agent retrieve current information at query time
  • Use web scraping to pull structured content from specific pages on demand
  • Restrict search to a whitelist of trusted domains to keep results focused and reliable
  • Configure timeout and path controls to manage how deeply pages are crawled

Channels & Handoff

Multi-channel integration
Reach Users Anywhere
Put your agent wherever your customers and teams already are.
  • Embed on any website or app with the <qlar-component> web component; inline or floating, designed in a visual builder with presets and a live preview
  • WhatsApp through Meta Cloud, Coster, or Qontak, with separate draft and live numbers
  • WhatsApp notification templates and an API to reach users first
  • Redirect to web: answer WhatsApp messages with a one-time sign-in link and continue on the web
  • Instagram messaging, plus sign-in with Google, Microsoft, or WhatsApp
Seamless live support
Escalation & Human Handoff
When the AI cannot resolve an issue, automatically route the conversation to a human agent without losing context.
  • Define trigger conditions, such as sentiment, topic, or explicit user request, that initiate escalation
  • Assign supervisors or support queues to receive escalated conversations by topic or area
  • Monitor a live queue of escalated chats and claim or release conversations in real time
  • Receive webhook events on escalation start, assignment, and resolution for external workflow integration

Automate & Extend

Scheduled & condition-based runs
New
Automations
Users ask in plain words; the agent sets up a task that runs on its own and reports back.
  • Created in conversation (chat, a voice call, or WhatsApp) and confirmed by the user before anything is saved
  • One-off or repeating schedules, always in an explicit time zone
  • Optional conditions checked against real data, so users only hear when something matters
  • Results arrive as a new conversation, a WhatsApp message, or both
  • An organization policy sets what's allowed, with limits, monitoring, and run history
Custom integrations
Use Your Own API
Connect your existing systems and services to automate workflows and deliver custom data in chat.
  • Import OpenAPI definitions or set up your own API endpoints
  • Configure actions such as HTTP method, parameters, and response handling, with no coding required
  • Authenticate with bearer tokens, basic auth, custom headers, or OAuth, and apply one setup to many APIs
  • Define how the call is handled, including call rules and response formatting, before it's shown to the user
Model Context Protocol
Connect to MCP Servers
Extend your agent with external tools and services using the Model Context Protocol standard.
  • Connect to any MCP-compatible server to give your agent new capabilities without building custom APIs
  • Enable or disable individual tools exposed by the MCP server to keep your agent focused
  • Authenticate per server: none, bearer token, custom header, server OAuth, or per-user OAuth consent
Ready-to-use apps
Use Built-in Plugins
The platform includes a set of built-in plugins: ready-made applications that eliminate the need to develop custom APIs.
  • SQL Database Reader: connect SQL Server, PostgreSQL, MySQL, or Oracle, even on-premise, through an outbound-only gateway, and let the agent answer in plain language, with row-level security
  • Chat Form: collect structured data from users mid-conversation using configurable question fields and conditional logic
  • Dynamic Table: let users query your data tables in natural language, with AI-assisted table creation and CSV/Excel import-export
  • Public Documents & Reminders: tap into a shared knowledge marketplace or schedule reminders, with an integration API and webhooks
  • Each plugin is configured through a simple UI, with no coding required
Seamless agent handoffs
Collaborate with Other Agents
Enable your agents to work together by securely calling one another for specialized tasks and seamless handoffs.
  • Agents work together automatically to decide what to do and when for more effective task handling
  • Agent keys control which agents may call yours
  • Each collaborator keeps its own ongoing thread, so context carries across multiple delegated calls, even during live voice conversations

AI Models & Cost Control

Model levels & BYO
Pick the Right AI Model, or Bring Your Own
Match the right model to the right task, on your terms.
  • Six built-in model levels (Light, Smart, Analyst, Insight, Thinker, and Sage), from everyday service to deep reasoning
  • "+" versions run the same model on Fast mode for quicker answers
  • Bring your own model and API key (OpenAI, Azure OpenAI, Azure AI, or DeepSeek); usage isn't credit-metered
  • Set organization-wide defaults per feature, override per agent, and cap response length
Manage AI expenses
Flexible Credit & Cost Control
Control token usage and model selection to keep your AI deployment cost-efficient at any scale.
  • Limit how much previous conversation the AI remembers to reduce token usage and save credits
  • Pick lightweight AI models to lower costs while still getting reliable results
  • Set maximum response lengths to avoid long, unnecessary replies that waste credits

Security, Privacy & Governance

Secure permissions
Access Control
Decide who can communicate with your agent and which information they are permitted to access.
  • Set the agent as public to increase reach or as invite-only to ensure exclusivity and secure critical contacts
  • Require sign-in before chatting
  • Create user groups (e.g., Finance, HR) so the agent knows what information can or cannot be shared with each team
  • Limit how often users can view source content to protect premium data and reduce overexposure
Keep only what you must
New
Privacy Mode & Sensitive Data
Decide what is kept, for how long, and what never enters a conversation.
  • Privacy Mode encrypts conversations and deletes them for good within 1–30 days; unreadable afterwards, even by Qlar
  • While it's on, no memory, query logs, or conversation analysis are kept
  • Document upload restrictions: describe what isn't allowed, such as ID or card numbers, and every upload is checked and refused with a reason
Multi-org, roles & audit
Run Multiple Organizations, One Team at a Time
Give every teammate the right level of access, and keep a full record of who changed what.
  • Create or join multiple organizations, each with its own agents, members, and shared credit pool
  • Assign roles (Owner, Admin, Builder, or Billing) so permissions match responsibility, or invite people with no role at all
  • Grant per-agent contributor rights (conversation history, personal data, automations), with every access logged
  • Transfer agents between organizations safely, and track a full audit trail of every change, down to the old and new value
Subscriptions & budget groups
Scale Spending With Confidence
From a first top-up to enterprise-wide budget controls, keep spend predictable at every stage.
  • Top up credits as you go, or subscribe to Starter, Pro, or Enterprise plans with automatic bonus credits
  • Set an auto top-up threshold and a monthly spending limit so you never run out, or overspend
  • Create budget groups to cap each user's monthly credit, and set a credit limit per agent
  • See exactly what each conversation cost, and export usage reports by agent, team, or time period

Test & Monitor

Confident releases
Test Automation
Build and run automated conversation test suites to verify your agent behaves correctly before and after every change.
  • Organize tests into scenario categories, such as onboarding, FAQs, or edge cases, for clear coverage tracking
  • Write conversation scenarios with pass/fail criteria to assert exact or semantic response correctness
  • Run the full test suite on demand and track results over time
  • Safely switch AI models by running the test suite against both versions before committing
Actionable insights
Analytics & Logs
Understand how your agent is performing and where it can improve through built-in dashboards and detailed logs.
  • View dashboard metrics including message volume, session length, resolution rate, and escalation frequency
  • Analyze user feedback signals to identify topics where the agent frequently falls short
  • Browse monitoring logs with service-level filtering to diagnose integration and tool-call failures
  • Export the dashboard, with all charts and metrics, as a PDF report
AI-powered insight at scale
Let AI Analyze Every Conversation
Go beyond dashboards: run AI evaluations across your entire conversation history to find what's really happening.
  • Run built-in analyses like complaint detection, sentiment, resolution score, and a 21-category risk assessment, or build your own
  • Analyze a full date range or a smart sample, and get aggregate charts plus AI-written insights automatically
  • Drill down from any chart straight into the conversations behind it, down to a single flagged message