Explorando las mejores aplicaciones de chatbots móviles: ¿hay una aplicación de chatbot gratuita para iPhone y Android?

Buscando aplicaciones de chatbot en 2026 generalmente significa una de dos cosas. Algunas personas quieren un asistente de IA personal como ChatGPT o Claude. La mayoría de los compradores de negocios quieren algo muy diferente: una aplicación que pueda responder a los clientes, dirigir leads, gestionar chat en vivo, manejar conversaciones de Messenger o Instagram, y permitir que un equipo se mantenga al día incluso cuando nadie está sentado en un escritorio.

Esta guía se mantiene en el lado empresarial de esa decisión. Si tu verdadero objetivo es un asistente de IA para consumidores, salta a nuestra guía sobre aplicaciones como ChatGPT. Si tu objetivo es la automatización orientada al cliente, el soporte, la captura de leads y la gestión de bandejas de entrada móviles, sigue leyendo. Esa es la versión del mercado que aún merece la frase aplicaciones de chatbot para la mayoría de los compradores.

Revisé páginas de precios oficiales, centros de ayuda y listados de aplicaciones móviles a partir del 12 de abril de 2026. La versión corta es simple. Tidio es la mejor opción general si tu sitio web es el centro de tu flujo de soporte. ManyChat sigue siendo la opción más fácil de pasar de gratuita a paga para DMs sociales. Freshchat es la aplicación de mesa de ayuda más limpia y económica para equipos que necesitan flujos de trabajo de agentes reales. HubSpot tiene más sentido cuando tu CRM ya está allí. Intercom es la opción de grado empresarial más conocida, pero rara vez es la más barata. Landbot es excelente para la calificación de leads web. Y si Facebook Messenger es donde ya aparecen tus leads y preguntas de clientes, MessengerBot es la lista corta más práctica que un conjunto genérico de chat en vivo.

No hay un único ganador porque estas herramientas están resolviendo diferentes trabajos. Esa es exactamente la razón por la que los rankings amplios se vuelven imprecisos. Una aplicación de DM social fuerte puede ser débil para el soporte de sitios web. Un conjunto de soporte empresarial pulido puede ser excesivo para un negocio de cinco personas. Un plan gratuito que se ve bien puede volverse costoso rápidamente una vez que entiendas el desencadenante de facturación. Si quieres el mapa de la plataforma más grande después de esta página, el comparativa completa de plataformas de chatbot es el seguimiento adecuado. Este artículo se mantiene más estrecho a propósito: las aplicaciones que los equipos reales pueden instalar, gestionar y justificar en este momento.

Cómo saber qué aplicaciones de chatbot realmente valen la pena usar en 2026

Una aplicación de chatbot es útil solo cuando hace una cosa específica mejor que tu configuración actual. Eso suena obvio, pero la mayoría de los errores de compra ocurren porque la gente salta esa prueba y va directamente a la marca, plantillas o cualquier característica de IA que tenga la página de aterrizaje más ruidosa.

El primer filtro es ajuste de canal. ¿Dónde comienzan tus conversaciones reales hoy? ¿Widget del sitio web, bandeja de entrada de la página de Facebook, DMs de Instagram, WhatsApp, correo electrónico o un portal de soporte? Si respondes a la mayoría de los compradores dentro de Messenger, un chatbot orientado a sitios web aún puede ser la compra incorrecta incluso si su demostración se ve más fuerte. Si tu carga de soporte es pesada en web y correo electrónico, un constructor de bots orientado a lo social puede crear más trabajo del que ahorra.

El segundo filtro es lógica de precios. En 2026, el software de chatbot ya no se precio de una manera clara. Algunas herramientas cobran por asientos. Algunas cobran por contactos activos. Algunas cobran por resoluciones de IA o cuotas de conversación separadas. Algunas parecen baratas hasta que tu equipo crece. Otras parecen caras hasta que te das cuenta de que reemplazan dos o tres herramientas por las que ya estabas pagando. La pregunta correcta no es solo, “¿Cuánto cuesta esta aplicación?” Es, “¿Qué evento exacto hace que mi factura aumente?”

El tercer filtro es la calidad de la transferencia. Muchas demostraciones de IA aún actúan como si el bot ideal es el que responde para siempre. En el uso real, el mejor bot es el que sabe cuándo dejar de pretender. Las disputas de reembolso, los casos inusuales, los clientes enojados y las conversaciones de ventas de alto valor necesitan un humano rápidamente. Si la aplicación hace que la transferencia sea torpe, eventualmente dañará el flujo de trabajo que prometió solucionar.

El cuarto filtro es la utilidad móvil. La palabra clave aquí es utilidad, no solo existencia en la tienda de aplicaciones. Muchos proveedores tienen una aplicación móvil. Mucho menos tienen un flujo de trabajo móvil en el que alguien de tu equipo realmente confiaría durante un viaje, en el piso de la tienda, entre reuniones con clientes o mientras cubre soporte fuera de horario. Esa diferencia importa más que capturas de pantalla brillantes.

El quinto filtro es la calidad del conocimiento. Las aplicaciones de chatbot de IA solo parecen inteligentes cuando el contenido detrás de ellas es claro. Lyro de Tidio, Breeze Customer Agent de HubSpot, Fin de Intercom y herramientas similares funcionan mejor cuando tu base de conocimientos, las páginas de tu sitio web y el material de preguntas frecuentes están actualizados y bien redactados. Si tu material de origen es vago, la aplicación responderá con confianza y aún así perderá el tiempo de todos.

El sexto filtro es qué tan rápido se vuelve útil la aplicación. Los equipos pequeños generalmente no necesitan otro sistema que tarde seis semanas en demostrar su eficacia. Necesitan un bot que pueda manejar un FAQ de las cinco principales, capturar un camino de leads, dirigir un tipo de problema de soporte o reservar un tipo de cita dentro de la primera semana. Por eso la velocidad de construcción sin código sigue siendo importante. Si aún estás en la etapa de diseño, nuestra guía para crear un chatbot sin código es la mejor lectura complementaria antes de elegir software.

Junta esos seis filtros y el mercado se vuelve más fácil de leer. ManyChat gana cuando los DMs sociales son el negocio. Tidio gana cuando el soporte en vivo del sitio web y la IA necesitan coexistir. Freshchat gana cuando necesitas un espacio de trabajo de soporte más limpio con un presupuesto más pequeño. HubSpot gana cuando el contexto del CRM importa tanto como el chat en sí. Intercom gana cuando necesitas profundidad de soporte empresarial y aceptas la facturación al estilo empresarial. Landbot gana cuando la conversación es realmente un formulario de leads disfrazado. MessengerBot gana cuando Facebook Messenger no es solo un canal adicional, sino el centro operativo de la conversación.

Best Chatbot Apps Compared Side by Side for Price, Free Plans, and Mobile Access

If you want the shortlist without the marketing spin, start here. I am not ranking these only by brand size. I am ranking them by how often they make sense for real businesses that need leads, support, or conversational automation to work this quarter, not someday.

The biggest thing that table shows is that chatbot apps are not one category anymore. Tidio and Freshchat are closer to support operations. ManyChat is closer to social growth and DM automation. HubSpot is a CRM-led service decision. Intercom is a help-desk and AI-operations stack. Landbot is part lead form, part conversational UX tool. MessengerBot is strongest when Facebook messaging is the real channel, not a side quest.

If I had to reduce the list even further, I would do it like this. Pick Tidio if your website is where most customer questions and pre-purchase objections land. Pick ManyChat if DMs on Instagram, Messenger, or creator-style social flows are where money actually happens. Pick Freshchat if you need a more structured support workspace without jumping to Intercom pricing. Pick HubSpot only if the CRM pull is already real. Pick Intercom if support is already a formal function, not a side responsibility. Pick Landbot if conversion on key web pages matters more than inbox handling. Pick MessengerBot if you already know Facebook Page conversations are where leads, bookings, or follow-ups live.

One more distinction matters here. If you are comparing apps because you need software to run the actual conversation flow, this list is the right place to start. If you are comparing apps because you still have not decided what kind of workflow to build, the smarter next read is the mejor chatbot para pequeñas empresas guide, because that article is organized around business fit first and vendor choice second.

Free Chatbot Apps You Can Test Before Paying for a Full Plan

Free plans are useful in this market, but not for the reason vendors want you to think. They are not there to run your entire business forever. They are there to answer one question quickly: will customers or leads actually use this conversation flow enough to justify paying for it?

ManyChat still gives the cleanest free starting point for social messaging. The new March 2, 2026 pricing model keeps a Free plan in place, and the paid ladder starts with Essential at $17 per month. That matters because you can test comment replies, story replies, basic DM automations, and simple lead capture without committing to a paid plan on day one. The main limitation is scale. Once engagement starts working, active-contact ceilings show up fast.

Tidio also handles the free-to-paid transition well, but in a different way. Its free version allows 50 billable conversations per month, up to 10 agents, and 100 flow visitors, while Lyro starts with a one-off 50 free AI conversations. That is enough to validate whether your site visitors will actually use chat and whether your FAQ content is good enough for AI to answer with confidence. It is not enough to support a busy store or a growing support queue for long.

Freshchat is one of the most underrated free options because it allows up to 10 agents. For small support teams, that is a serious test environment, not a toy. The catch is that the richer omnichannel and AI layers live in the paid plans. So the free tier is best treated as a way to prove workflow fit, not as the final destination.

HubSpot has the classic CRM-style free entry point. The free tools are useful if you already want basic inbox, contact, and service functionality tied together. The problem is that businesses sometimes mistake “free” for “good enough.” In practice, the reason to start free on HubSpot is to validate process and data structure, not to avoid paying forever for a customer-service operation.

Landbot is free in the sandbox sense. Its free-forever plan gives you a real place to prototype lead flows, test web questions, and see whether your conversational form logic makes sense. It is excellent for proof of concept. It is not the best final answer once traffic becomes real.

Free trials matter just as much as free plans. MessengerBot, Intercom, and several others use trials rather than permanent free tiers because they expect real businesses to outgrow the free layer quickly. That is not automatically worse. In fact, it is often more honest. A free trial that gives you realistic functionality for two weeks can be more valuable than a forever-free plan that breaks the moment the workflow starts helping.

The right way to use free chatbot apps is simple:

• Test one high-frequency question or path first. FAQs, lead capture, quote request, booking, or order status are ideal.
• Measure real usage, not setup excitement. People launching a bot always overestimate how much the first build proves.
• Watch what fails first. If limits, handoff, or knowledge quality break immediately, that tells you more than the homepage copy.
• Upgrade only when the free tier is now costing time or revenue. That is the real signal.

If you need ideas for the first workflow worth testing, the best companion piece is this roundup of chatbot use cases that generate revenue. The free plan is only valuable when the use case itself is worth the setup time.

Chatbot Apps That Work Best When You Need to Reply From Your Phone

Not every chatbot app with a mobile presence is actually useful from a phone. That distinction matters because a lot of teams do not manage conversations from a single desktop help-desk shift anymore. They reply between meetings, while traveling, from a retail counter, or during after-hours coverage.

ManyChat is one of the strongest phone-first tools here. Its current iPhone app has a 4.5 rating from 1.9K ratings, and the positioning is clear: manage DMs, launch flows, handle engagement, and turn comments into customer conversations directly from your pocket. That maps well to the product’s real strength. If your conversations start on Instagram, Facebook Messenger, WhatsApp, or Telegram, ManyChat still feels built around mobile reality rather than desktop nostalgia.

Tidio is the strongest all-around mobile support app in this list. The iPhone app currently sits at 4.8 from 4.8K ratings, and the feature pitch lines up with what buyers need: live chat, chatbots, quick responses, visitor details, and push notifications. More important than the rating is the workflow. Tidio lets a small team stay responsive without carrying a full enterprise support stack everywhere.

HubSpot is the best mobile choice when the conversation is inseparable from the CRM. Its iPhone app is currently rated 4.7 from 15K ratings, and that scale matters. You are not only seeing the message thread. You are seeing the record, history, and commercial context behind it. That makes HubSpot much stronger for sales follow-up and customer management on the go than a lighter-weight chat app that only shows the inbox.

Freshchat is usable on mobile, but it feels more agent-centric than growth-centric. Its iPhone app currently shows 3.6 from 51 ratings, which is not disastrous, but it is not a dominance signal either. The app makes sense for teams that want to view, assign, and manage conversations from anywhere. It is less compelling if you expect a slick creator-style or social-growth mobile experience.

Intercom is the clearest example of why app-store presence and mobile quality are not the same thing. The Intercom Conversations app currently sits at 2.3 from 144 ratings on the U.S. App Store. That does not make Intercom a bad platform overall. It does mean phone-first teams should be realistic: Intercom remains stronger as a broader support platform than as a beloved mobile inbox experience.

This is where buying context matters. If your team lives on phones and social channels, ManyChat is usually the easier fit. If your team lives on phones and website support, Tidio is stronger. If your reps need CRM context more than chat-only speed, HubSpot wins. If your team already works in a structured help desk and just needs emergency mobile coverage, Freshchat and Intercom can still work, but neither is the cleanest mobile-first experience in the group.

The practical rule is this: if phone handling is core to the workflow, do not treat the mobile app as an afterthought. Install it before you buy. Reply to real messages on it. Reassign a conversation. Check whether saved replies, contact history, notes, and handoffs feel natural. That ten-minute test tells you more than any feature list.

Best Chatbot Apps for Small Businesses, Ecommerce Stores, and Lead Capture

Most small businesses do not need the most advanced chatbot app. They need the one that maps cleanly to where buying conversations already happen. That means the shortlist should be organized by commercial situation, not by which vendor used the word AI most aggressively.

If Facebook Messenger is the real sales or support channel, MessengerBot is the strongest fit. The reason is not abstract AI quality. It is operational fit. MessengerBot is built for Page-level conversation flows, lead capture, automations, website chat extensions, and follow-up logic without pushing you into active-contact billing every time engagement improves. If that sounds like your setup, Verificar precios actuales before you assume a broader live-chat platform is the smarter buy.

If the business runs on Instagram, comments, story replies, or creator-style DM funnels, ManyChat stays near the top. Its March 2026 pricing model made the path easier to understand: Essential starts at $17 per month, Pro at $39, and Business at $99, with active-contact allowances climbing as you scale. That is still not “cheap forever,” but it is much clearer than older versions of the ManyChat conversation. If your revenue path starts with social engagement, it is still one of the fastest tools to turn attention into reply chains and captured leads.

If your website is where pre-purchase questions and support volume pile up, Tidio is the safest all-around choice. It combines live chat, ticketing, and AI agent features in a way that makes sense for ecommerce and service businesses alike. The entry pricing is reasonable, the free layer is good enough to test seriously, and the mobile support story is much stronger than many heavier support stacks.

If your main job is lead qualification on the site itself, Landbot is stronger than a standard support widget. This is the tool you use when the conversation is really a guided form, quote path, or product-fit quiz. It is less ideal as a general-purpose support desk, but extremely strong when the goal is to reduce friction between curiosity and conversion on a specific page.

If your team already lives inside HubSpot, adding another standalone chatbot app is often the wrong move. HubSpot’s value is not just chat. It is the shared customer context. A chatbot connected to the CRM, tickets, deals, and lifecycle stages is much more useful than a separate app pretending it can infer that context from one conversation. That is why HubSpot is a better buy for CRM-native teams than for businesses that only need a lightweight chat tool.

For ecommerce specifically, the best app depends on the store’s bottleneck. Use Tidio when the problem is high-volume website support. Use ManyChat when the problem is social selling and DM-based conversion. Use Landbot when product selection, lead qualification, or quote capture is the real web bottleneck. Use MessengerBot when Facebook is still where shoppers start the conversation. If you want to pressure-test your shortlist against actual deployment patterns, this guide to the mejor chatbot para pequeñas empresas is the right next read.

The biggest mistake small businesses make here is buying for possibility instead of workload. A sophisticated support suite looks impressive, but if you only need to answer 200 repetitive messages a month and catch a handful of missed leads after hours, it may be slower and more expensive than a simpler app built for your actual channel mix.

AI Chatbot Apps That Reduce Customer Support Load Instead of Adding More Work

This is the part of the market that changed fastest. In 2024 and 2025, AI chat often felt like a promising extra. In 2026, the leading chatbot apps are openly selling AI resolution as the main value layer. That does not mean every claim is universal. Most of the best performance numbers are vendor-reported. But it does mean the market is mature enough that AI is no longer just a decorative feature in the plan comparison table.

Tidio says Lyro resolves 67% of requests. HubSpot says Breeze Customer Agent already resolves 65% of conversations, cuts resolution time by 39%, and is active across more than 8,000 customers. Intercom now prices Fin at $0.99 per outcome across its core support plans, which only makes sense because Intercom believes buyers will accept resolution-based pricing if the automation is actually useful. Those are not small signals. They show where the category is going.

The practical lesson is not that every team should switch on AI immediately. It is that a chatbot app now has to be judged on three support questions at once:

• Can it answer simple questions well enough to deflect repeat work?
• Can it hand off cleanly when it should?
• Can your team afford the billing model if the AI actually succeeds?

Tidio wins a lot of SMB support comparisons because it keeps those three questions relatively balanced. You can test Lyro with 50 free conversations, see whether your knowledge base is strong enough, and then decide whether the AI layer deserves a budget. That is much easier to justify than an enterprise contract before you have any proof.

HubSpot is more interesting than older chatbot buyers sometimes realize. HubSpot’s own knowledge-base docs show Customer Agent can be assigned to live chat, Facebook, and WhatsApp channels, and it can route to humans when confidence is low or escalation rules are met. That is powerful because the AI is not floating alone. It is attached to the CRM and the help desk. The caveat is that the AI layer only makes commercial sense if you already want the surrounding HubSpot system.

Intercom is the strongest fit when support is already a real operation and not just “whoever is around answers the chat.” Fin is included across Intercom’s plan structure, but the usage-based pricing makes the economics visible. That transparency is good. It also forces discipline. If you buy Intercom, you should know how many conversations you expect Fin to resolve and what those outcomes are worth to the business.

Freshchat belongs in this section too because it provides a more affordable path into structured support. The paid plans include Freddy AI Agent eligibility with the first 500 sessions included, which is a good on-ramp for teams that want to experiment with AI support without jumping into the most expensive tier of the market.

The support buying mistake I keep seeing is this: teams compare AI chatbot apps as if the AI alone decides the outcome. It does not. Source quality, escalation rules, inbox ownership, analytics, and pricing model matter just as much. If your FAQ content is weak, no AI layer will save it. If your team does not know who owns escalations, the bot will just delay frustration. If you want the deeper budget view after this section, the guía de precios del chatbot breaks down the exact billing patterns that start looking cheap and finish looking dangerous.

Real Chatbot App Pricing in 2026: What You Pay for Seats, Contacts, and AI

The hardest part of choosing between chatbot apps is not feature comparison. It is understanding what you are actually buying. The numbers below reflect public pricing and help-center details a partir del 12 de abril de 2026, and that date matters because pricing is still moving fast in this category.

ManyChat is the clearest example of a pricing model that looks inexpensive until engagement starts working. On the new March 2, 2026 system, Essential includes up to 250 active contacts, Pro includes 2,500, and Business includes 7,500. That is sensible. It is also the reason buyers need to think in engaged contacts, not just follower counts. A good campaign can raise the bill faster than expected.

Tidio is trickier because the core support plan and the AI layer are related but not identical. You can start with the free and Starter layers, but the real cost story becomes clearer once you decide how many Lyro conversations you want each month. That does not make Tidio bad value. It just means the plan badge is not the full budget number.

Freshchat is easier to explain: free for up to 10 agents, then Growth at $19 per agent, Pro at $49, and Enterprise at $79, all billed annually on the public pricing page. Freddy AI Agent includes the first 500 sessions on paid plans, which is generous enough for evaluation but not enough to ignore the AI layer forever if volume is real.

HubSpot needs one exact date callout because this is where older screenshots go wrong. HubSpot announced last week that Breeze Customer Agent moves from $1.00 per conversation to $0.50 per resolved conversation, with the change taking effect on 14 de abril de 2026. So if you are reading this a partir del 12 de abril de 2026, you are in the brief window where the live pricing story includes both the current Service Hub seat pricing and the announced near-term change in how Customer Agent will be billed.

Intercom is the cleanest high-end pricing story. Essential is $29 per seat annually, Advanced is $85, Expert is $132, and Fin is $0.99 per outcome. There is no magic here. If Fin solves a lot, you pay for that. For serious support teams, that transparency is useful. For small businesses that mainly need basic automation and a few fast handoffs, it can be more than they need.

The best pricing model depends on what kind of headache you prefer. Flat pricing is easiest to forecast. Contact-based pricing can feel friendlier until audience engagement spikes. Outcome-based pricing feels fair until the AI gets good. Seat-based pricing is predictable until more teams want access. That is why the smartest buyers map pricing to workflow first, not to brand preference.

How to Choose the Right Chatbot App Without Overbuying Features You Will Not Use

The fastest clean decision is to stop asking which chatbot app is the most advanced and start asking which one matches the first problem you need solved. If the answer is “I am not sure,” you are not ready to compare vendors yet. You are still choosing the job.

1. Pick the primary channel first. Website, Facebook Messenger, Instagram, WhatsApp, or help desk. If you get this wrong, the rest of the shortlist will be wrong too.
2. Name the first workflow in plain language. FAQ deflection, booking, quote capture, comment-to-DM, order status, or lead qualification. One narrow goal beats a vague “AI support” ambition.
3. Decide whether mobile handling is optional or critical. If people on your team will reply from phones often, test the mobile app before you pay.
4. Find the billing trigger. Seats, contacts, conversations, or outcomes. Do not buy until you know which behavior raises the bill.
5. Check how the app handles human handoff. This is where a lot of demos hide the real operational pain.
6. Look at one real integration you need now, not ten you may want later. CRM, Google Sheets, ecommerce, calendar, or messaging channel. Start with the blocker, not the fantasy stack.
7. Run a two-week test with a real use case. A serious chatbot app should prove something useful in that window.

If you want a quick buyer map, use this shortcut:

• Choose MessengerBot if Facebook Messenger is where the business already lives.
• Choose ManyChat if social DMs and creator-style automations are the commercial engine.
• Choose Tidio if you need website support plus AI without enterprise complexity.
• Choose Freshchat if you need a support workspace first and AI second.
• Choose HubSpot if the CRM context matters as much as the chat widget.
• Choose Intercom if you run a serious support function and can justify usage-based AI spend.
• Choose Landbot if the conversation is really a guided web conversion flow.

The second shortcut is even simpler. If the idea of building the first bot still feels fuzzy, do not over-research software. Go build a first narrow workflow instead. That is where how to build a chatbot without code becomes more helpful than another pricing chart. The market gets much easier to read once you know whether you need a lead bot, a support bot, or a messaging bot.

Common Chatbot App Mistakes That Kill Adoption, Lead Quality, and ROI

The first mistake is buying by AI hype instead of by workflow. A lot of businesses now say they want an AI chatbot app when what they really want is one of three things: fewer repetitive support tickets, faster lead capture, or a cleaner mobile inbox. If the tool does not improve one of those fast, the AI label will not save it.

The second mistake is underestimating channel gravity. Teams often buy a website chat app because it looks polished, then discover that most real customers still start in Facebook or Instagram messages. Or they buy a social automation tool and realize the real bottleneck is not DMs at all. It is order-status questions on the site. The app is only as good as its match to where the customer already is.

The third mistake is treating the free plan like a long-term strategy. Free plans are great for validation. They are bad at carrying a workflow that is already proving value. If the bot is already booking jobs, qualifying leads, or deflecting tickets, refusing to pay for the right tier is often the more expensive choice because the limits start shaping behavior.

The fourth mistake is not testing the mobile workflow with real conversations. This matters more than teams expect. ManyChat, Tidio, HubSpot, Freshchat, and Intercom all look acceptable on a feature matrix. They do not feel equally good when you actually try to clear live conversations from a phone. If your team works on the move, that difference becomes operational, not cosmetic.

The fifth mistake is forgetting that AI quality is mostly source quality in disguise. Tidio, HubSpot, Intercom, and similar AI layers perform far better when the knowledge base is clean, current, and written the way customers actually ask questions. Bad FAQ content creates bad AI answers. That is not a model failure. It is a content failure.

The sixth mistake is skipping human ownership. A chatbot app should not create a mystery zone where nobody knows who receives escalations, who updates knowledge, or who reviews conversations that the bot handled badly. The teams that get ROI usually assign one owner early, even if the tool itself is shared across support, marketing, and sales.

The seventh mistake is comparing consumer AI assistants to business chatbot apps as if they solve the same problem. They do not. ChatGPT can help a human draft replies. It cannot replace a purpose-built app that manages channels, contact rules, routing, handoff, reporting, and inbox operations. That is why this article stays business-first. If you want the broader “what can a bot actually do for revenue?” angle after this, revisit the revenue-focused chatbot use cases before you buy software for a problem you have not defined.