Questions, answered plainly.

Your visitor's question is matched against your Q&A pairs. Close enough match → the model answers grounded in that content. No close match → it says it doesn't know and offers to escalate, rather than guessing.

Minutes, not a sprint. Write a handful of Q&A pairs in the dashboard, paste one script tag into your site. No framework or backend integration required.

No — signup and the dashboard are open immediately. Email verification only gates revealing the widget install snippet (Settings → API key, or step 3 of onboarding), so a working chat widget can't end up live on a site before we've confirmed the signup email is real.

Not yet. Today the model's instructions are fixed to strict grounding. Configurable tone and escalation rules are a real gap, not a shipped feature.

The bot tells them to when it can't answer. Routing that into a live agent queue or ticketing system isn't built yet.

No — writing Q&A pairs is plain text, and embedding is one script tag. No API integration is needed for basic use.

Yes — the dashboard shows Q&A coverage, message volume, and a real count of questions it couldn't answer, so gaps in your knowledge base are visible, not hidden.

Yes — every tenant's knowledge base, conversations, and API key are scoped at the database layer. Nothing is shared or mixed across accounts.

Plan tiers by Q&A pair count and messages per month, with a genuinely free tier to start. See the Pricing page for exact numbers and prices in your currency.

Email convo.rail.ai@gmail.com — there's no live chat or ticketing system on our own side yet, so email is the direct line to a real person.

Starter includes 2 seats, Growth 3, Scale 4 (Free is solo-only, 1 seat). Need more? Buy extra seats anytime from Settings > Team in your dashboard - bundles start at ₹299 for 1 seat, down to ₹1,499 for 10 (larger bundles cost less per seat), charged instantly via Razorpay.

Sign up - no credit card required - and you're in the dashboard immediately, ready to add your first Q&A pairs.

The dashboard is open right after signup, but the install snippet (your API key) only unlocks after email verification - so a working widget can't go live on a real site before we've confirmed the signup email is real.

Add a handful of Q&A pairs covering the questions your visitors actually ask, before installing the widget on your live site.

Minutes, not a sprint - write a few Q&A pairs, paste one script tag.

No - writing Q&A pairs is plain text, and installing the widget is pasting one script tag. No coding knowledge required for the basic setup.

Yes - the dashboard is accessible right after signup. Add your real Q&A pairs whenever you're ready.

The bot has nothing grounded to answer from, so it'll decline almost every question. Add your Q&A pairs first so visitors get useful answers from day one.

Onboarding today is the dashboard plus the Live Demo page to try things out - not a separate multi-step guided wizard.

Multi-user team accounts aren't a documented feature today - one login per account.

No - the Free plan requires no credit card.

Yes - the Live Demo page runs a real, working widget you can test before touching your own site.

Start with the questions your visitors actually ask most often, phrased the way they'd ask them - not how you'd phrase it internally.

There's no required minimum - the Free plan supports up to 5, so even a small starting set is enough to go live and expand from there.

No documented bulk-import option today - Q&A pairs are added manually in the dashboard.

That your Q&A pairs cover your common questions, that you've tested a few in the dashboard or Demo page, and that your API key is the one actually in your script tag.

One script tag and an API key from your dashboard. No SDK, no backend integration, no framework requirement for the basic setup.

Not for the basic install - pasting a script tag is something most site builders and CMS platforms support without code.

Any site that lets you paste a script tag - WordPress, Shopify, Webflow, Wix, Squarespace, or a fully custom-coded site. It's plain JavaScript, not a platform-specific plugin, so it works the same way everywhere.

As a chat bubble your visitors click to open, the same pattern used on the Demo page - it stays out of the way until someone needs it.

Not as a self-serve option today - widget theming isn't exposed in the dashboard yet.

It's built as a lightweight, framework-independent script rather than a heavy embedded app, so it stays light on your site regardless of what you built your site with.

Each account gets one API key. Using it on more than one domain isn't actively blocked, but it isn't a documented, supported multi-site feature yet - email us if this is what you need.

Yes - it's a standard responsive web widget that runs wherever the script loads, including mobile browsers.

Yes, any time from the dashboard - useful if a key is ever compromised.

No - once embedded, the widget runs independently on your site. The dashboard is only for managing Q&A content and viewing usage.

Not published today - ConvoRail is currently used through the dashboard and the embeddable widget script.

That depends on how you place the script tag - add it sitewide via your CMS's global footer, or only to specific page templates to limit where it appears.

No documented third-party integrations today beyond the embeddable widget itself.

It's built and documented as a website widget today - not a mobile-app SDK.

Yes, any time - changes in the dashboard reflect on the live widget without needing to reinstall anything.

Yes - Free is a real plan, not a time-limited trial: 5 Q&A pairs and 200 messages/month, with the ConvoRail badge shown. No credit card required.

Free, Starter, Growth, and Scale - plus a custom Enterprise tier for businesses that need something beyond the standard limits.

Q&A pairs are the size of your knowledge base - how many question-and-answer entries you can have. Messages are how many chat replies your bot sends in a month.

The same grounded chat engine, usage dashboard, one-line embed, and API key access on every tier - only Q&A pair count, monthly messages, and widget branding differ by plan.

Plans gate on Q&A pair count and monthly messages the bot actually answers - not raw site visitors. A visitor who never opens the chat doesn't count against your plan.

Only on the Scale plan. Free, Starter, and Growth all show the badge - removing it is a Scale-tier differentiator.

A custom Q&A pair limit, custom message volume, and dedicated onboarding and support. It's priced individually - contact sales rather than self-serve signup.

No separate line item - the grounded chat engine, including every answer Anok AI gives, is included in your plan's price and gated only by your monthly message limit.

Compare your expected monthly question volume and knowledge-base size against the limits on the Pricing page - start on Free or Starter and move up as you outgrow the caps.

Yes - your dashboard's usage view shows Q&A pairs used and messages sent against your plan's cap.

You won't be able to add new pairs until you either remove existing ones or move to a higher tier.

Yes - prices are localized automatically based on your detected location, falling back to a global USD price where we don't have a specific rate.

No - it's marketed as an ongoing real plan, not a time-limited trial.

No documented add-on or top-up option today - moving to a higher tier is the current path for more headroom.

Per month, as shown on the Pricing page.

Billing and subscriptions run through Razorpay.

Prices are localized automatically based on your detected location. If we can't detect it, or your country doesn't have a specific rate set, you'll see the global USD price.

No. No setup fee, no surprise per-message add-on charges - what's shown on the Pricing page for your plan is what you pay. Going over your plan's limits caps usage rather than silently billing you more.

Usage is capped, not silently overcharged - once you hit your plan's limit, you'll need to move to a higher tier to keep going, rather than getting an unexpected bill.

No - the monthly message allowance resets each billing period rather than accumulating.

Self-serve cancellation isn't wired into the dashboard yet - email convo.rail.ai@gmail.com and we'll handle it manually in the meantime.

Not formally published yet - ConvoRail's Terms of Service are still a pre-launch draft. Reach out directly if you have a billing concern and we'll sort it out.

No long-term contract has been published - plans are shown and billed on a monthly basis.

Self-serve upgrades and downgrades aren't live in the dashboard yet - email us and we'll move your account manually.

Not formally documented yet - email us before downgrading or cancelling if you have questions about your account's data and we'll clarify for your specific case.

Not documented in detail today - ask via email if you need a specific invoice for your records.

That's what the Enterprise plan is for - contact sales to discuss a custom arrangement.

Not specified on the Pricing page today - tax handling depends on your region at checkout.

convorail.com serves markets outside India; convorail.in is the dedicated India-market site. Same company, billing handled per site.

Not published today - plans are shown at their monthly rate. Ask via email if annual billing would help your situation.

It's a Q&A knowledge base, a grounded chat engine, a one-line embeddable widget, and a usage dashboard - see exactly what ConvoRail does on the Features page.

Every reply is checked against your Q&A pairs before it's sent - if nothing matches closely enough, it says it doesn't know instead of improvising an answer.

Q&A pair coverage, message volume against your plan's limit, and how many questions the bot couldn't answer.

Yes - unanswered-question volume is visible in the dashboard, so gaps in your knowledge base show up instead of getting silently swallowed.

Yes - every tenant's knowledge base, conversations, and API key are scoped at the database layer. Nothing is shared or mixed across accounts.

No - the dashboard is web-based today, no dedicated mobile app.

No documented rich-media reply formatting beyond text today.

No documented organizational structure beyond a flat list of Q&A pairs today.

No documented duplicate-detection feature today.

No documented export option today.

The dashboard covers Q&A coverage, message volume, and unanswered-question counts - no deeper analytics (like top-asked questions) documented beyond that today.

The Live Demo page and dashboard let you test questions before going live - there's no separate formal "staging" toggle for your own widget today.

Configurable tone and escalation rules, routing to a live human agent, and compliance certifications are all explicitly flagged as roadmap items, not shipped features.

No - it's one codebase for every tier. Plans only gate Q&A pair count, monthly messages, and widget branding, not which features you get.

Anywhere a feature isn't built yet, the relevant FAQ answer or page copy says so directly rather than implying it's already available.

Anok AI is the name of the conversational agent that powers every ConvoRail chat - the "who" behind the widget when your visitors ask a question.

No - Anok AI is the agent built into ConvoRail. You don't sign up for it separately; it's what runs every time your widget answers a question.

Your visitor's question is matched against the Q&A pairs in your knowledge base. On a close-enough match, Anok AI answers grounded in that content - not from general internet knowledge.

By design. If nothing in your knowledge base closely matches the question, Anok AI declines and signals a handoff instead of guessing at an answer that could be wrong.

Preventing exactly that is the core design goal - grounding restricts it to your Q&A pairs rather than open-ended generation. Like any AI system this isn't a mathematical guarantee, but it's what the product is built around.

OpenAI's models generate the grounded reply, per ConvoRail's Privacy Policy.

Within an active session, recent turns are taken into account for a limited window. It isn't built as a long-term, persistent memory of a visitor across separate visits.

It answers in whatever language your Q&A pairs and the visitor's question use, since the underlying model generally understands multiple languages - though this isn't a separately tested or guaranteed multilingual feature today.

Not as a configurable option today - its tone is fixed to the strict-grounding behavior. Configurable tone is a real gap, not a shipped feature.

No - it answers questions from your Q&A pairs. It doesn't perform transactions, bookings, or account changes.

The engine and grounding behavior are the same for everyone. What it actually knows differs entirely, based on each business's own Q&A pairs.

It replies in real time, at the underlying model's typical response speed. There's no published response-time guarantee or SLA today.

Yes - your dashboard's usage view shows unanswered-question volume, so gaps in your knowledge base stay visible instead of silently swallowed.

Not automatically - it only improves as you add or refine your Q&A pairs. There's no self-learning from past conversations today.

It tells the visitor it doesn't know and signals that a handoff is needed. Routing that into a live-agent queue or ticketing system isn't built yet.

Email convo.rail.ai@gmail.com - that's the direct line to a real person today.

No separate live-support widget for ConvoRail's own customers is documented today - email is the current channel.

Not published today - no formal support-hours commitment yet.

This FAQ page is the current self-serve resource - no separate help-center product exists today.

No published response-time guarantee today.

Anok AI tells them it doesn't know and signals that a handoff is needed - routing that into an actual live-agent queue isn't built yet.

Usually that means a Q&A pair needs editing on your end - update it directly in the dashboard. If it seems like an actual product bug rather than a content gap, email us.

Not documented - email is the confirmed support channel today.

Yes - dedicated onboarding and support is part of the Enterprise plan.

Email with a description of what happened - there's no separate public bug tracker or support portal today.

Yes, by email - there's no public roadmap-voting tool today.

Not documented today.

As a pre-launch product, ConvoRail doesn't publish an uptime guarantee or SLA today - see the Terms page for the full framing.

Enterprise customers get dedicated onboarding help; other tiers can email for guidance while self-serve onboarding tooling is still basic.

The same address, convo.rail.ai@gmail.com, handles both today.

An AI chatbot for your website that answers strictly from a knowledge base you write yourself, powered by Anok AI - the conversational agent behind every chat.

Sakshi Sidana, Founder & CEO, with a background in brand and marketing strategy. See the Founder page for her story.

Jaipur, Rajasthan, India.

Any business that gets repeat questions on its website - SaaS, e-commerce, professional services, or local businesses - and wants those answered accurately without someone monitoring a chat window all day.

Mayank Sachdev (Marketing Head), Ayush Arora (Sales Head), and Kanika Sachdev (Ops Head), alongside founder Sakshi Sidana. See the About page for the full team.

convorail.com serves every market outside India; convorail.in is the dedicated site for the India market. Same company, same product, split by geography.

No, deliberately not. It only answers from the Q&A pairs you give it - a narrower job than a general assistant, on purpose, since that's what makes the grounded-answer behavior possible.

Most competing tools default to sounding confident even when they're not grounded in anything you wrote. ConvoRail's default is the opposite: no close match in your Q&A pairs means a stated refusal, not an improvised answer.

Building a chatbot that's honest about what it doesn't know, so businesses can trust automated answers instead of policing them for mistakes.

No funding information is published today.

No specific founding date is published - the product is early-stage and pre-launch today.

convo.rail.ai@gmail.com for product questions and support.

Signup and the dashboard are open now. The product is still early, and features that aren't built yet are explicitly marked as roadmap rather than shipped throughout this site.

Yes - Instagram, LinkedIn, X (Twitter), Pinterest, and Threads. Links are in the site footer.

Open roles are listed on the Careers page when they exist. If you don't see a fit but think you'd be a great addition anyway, email us.

Yes. When the bot can't answer, it can show an in-chat contact form so the visitor leaves their details, and the conversation is flagged for your team in the inbox.

Yes. The chat widget lets visitors rate a conversation, and the rating is shown to your team in the inbox so you can see where answers fell short.

Today ConvoRail covers website chat and Gmail in one inbox. Channels such as WhatsApp, Facebook Messenger, Instagram and Outlook are on the roadmap and are not presented as live.

Yes. You can start on the free plan with no credit card, add your own Q&A pairs and see how the chatbot answers before choosing a paid plan.