Question 1

What is Verly?

Accepted Answer

Verly is a developer-first platform for building and deploying omni-channel AI agents. We provide the infrastructure (LLM orchestration, RAG, telephony, and state management) so you can focus on building great support experiences via code or our config UI.

Question 2

Who is Verly for?

Accepted Answer

We are built for engineering and product teams who want full control over their AI support stack. Unlike black-box chatbot tools, Verly offers comprehensive SDKs, webhooks, and observability to integrate deeply with your existing backend.

Question 3

Which LLMs do you support?

Accepted Answer

We are model-agnostic. You can bring your own API keys for OpenAI (GPT-4), Anthropic (Claude), or use open-source models hosted on Groq/Together AI for ultra-low latency. We also handle model fallback and context window management automatically.

Question 4

Can I train the chatbot with raw text (no pages/files)?

Accepted Answer

Yes — paste raw text into your knowledge base and we’ll index it the same way as scraped pages or uploaded docs. This is perfect for internal SOPs, policies, and “things that aren’t on your website”.

Question 5

What content can I use to train the chatbot?

Accepted Answer

Website URLs (single page or full site), sitemaps, docs/knowledge bases, and uploads like PDF/DOCX/TXT/MD. Mix sources — the bot learns better with broader coverage and fewer gaps.

Question 6

Do you automatically re-train when my website changes?

Accepted Answer

If you enable auto-sync, we periodically re-crawl/re-index your sources so the bot stays up to date. You can also trigger a manual sync anytime when you ship a big update.

Question 7

What happens when the bot can’t answer?

Accepted Answer

You can configure escalation rules so the bot hands off to a human, captures contact info, creates a ticket, or routes the conversation to a team inbox — while preserving full conversation context.

Question 8

How long does training take?

Accepted Answer

Usually minutes. It depends on the total content volume and the number of sources. You can start testing immediately as new content is indexed.

Question 9

How do I integrate Verly with my database?

Accepted Answer

Agents can trigger 'Actions' (function calling) to fetch data or perform tasks in your system. simply define your API schema, and the agent will intelligently query your backend to check order status, book appointments, or update user records.

Question 10

Can I deploy to WhatsApp and Web simultaneously?

Accepted Answer

Yes. You define your agent's behavior once—prompts, tools, and knowledge base—and deploy it across multiple channels (Web Widget, WhatsApp, Twilio Voice) instantly. We handle the channel-specific protocol conversion.

Question 11

What is the latency for Voice Agents?

Accepted Answer

Our voice infrastructure is optimized for real-time conversation. By using streaming transcription and edge caching, we achieve varying latencies as low as 500ms, providing a natural 'interruptible' conversation experience.

Question 12

Can I embed the chatbot on my website?

Accepted Answer

Yes. You can embed it with a snippet and customize branding, colors, position, and behavior. You can also link directly to a hosted chat page.

Question 13

Can I integrate with Slack/Zendesk/CRM tools?

Accepted Answer

Yes — integrate via API/webhooks and connect support workflows (handoff, ticket creation, lead capture). If you tell us your stack, we’ll recommend the cleanest integration path.

Question 14

Is the widget customizable / white-label?

Accepted Answer

Yes. You can match your site’s design and optionally remove branding depending on your plan. You can also tune tone, model behavior, and escalation triggers.

Question 15

Is my data used to train models?

Accepted Answer

No. Verly offers strict data isolation. We do not use your customer conversations to train our foundational models, and we support enterprise agreements with zero-retention policies for sensitive industries like Healthcare and Finance.

Question 16

How does pricing work?

Accepted Answer

We offer a usage-based pricing model. You pay for what you use—per conversation minute for voice, and per message for chat. There are no mandatory upfront contracts for our Starter tier.

Question 17

Do you offer on-premise deployment?

Accepted Answer

Yes, for Enterprise customers, Verly can be deployed within your own VPC (AWS/GCP/Azure) or creating a dedicated instance to ensure maximum compliance and security.

Question 18

How does the free trial work?

Accepted Answer

You get full access for the trial period (no long-term commitment). When the trial ends, you can upgrade to keep the bot live or stay on a limited/free tier if available.

Question 19

What counts toward usage limits?

Accepted Answer

Typically: message volume, indexed content size, and/or number of chatbots/workspaces. If you’re unsure which lever matters most for you, ping support and we’ll map your expected traffic to the right plan.

Question 20

Do you offer plans for agencies?

Accepted Answer

Yes — if you’re building bots for clients, we can set you up with an agency-style workflow (multiple bots, separate client environments, and billing options). Email us and we’ll get you the details.

Everything you need to know before you launch.

General & Platform

Chatbot Training and Support

Integration & Deployment

Security & Pricing

General & Platform

Chatbot Training and Support

Integration & Deployment

Security & Pricing

Ask your favorite AI assistant about Verly.

Web