natural-sounding ai receptionist with a personality of its own

natural-sounding ai receptionist with a personality of its own

handle business enquiries and take messages with an ai receptionist that follows your instructions

see what it's like to talk to a voice agent

see what it's like to talk to a voice agent

see what it's like to talk to a voice agent

…

these are demo agents. they can't take real actions or give real support, so please don't share personal details.

clone this agent ↗

leave the everyday calls to your ai receptionist agent

leave the everyday calls to your ai receptionist agent

leave the everyday calls to your ai receptionist agent

shape each conversation around the reason someone called and what they need next

  • business enquiries

    an ai receptionist answers questions using the business information provided and explains what callers need to know about a service before visiting

  • appointment requests

    asks what callers want to book and captures their preferred time as an appointment request for the team to confirm

  • messages and callbacks

    takes a message for the person the caller wants to reach and collects their contact details along with the reason for a callback

set up your ai receptionist agent in three steps

set up your ai receptionist agent in three steps

set up your ai receptionist agent in three steps

start with a template or build from scratch around the conversations the agent needs to handle

  1. choose from 8 agent templates

    pick a template for the work you want the agent to handle, or build your own from scratch.

  2. configure the prompt and voice

    set the agent’s role in its system prompt and choose from 4 voices for its calls.

  3. deploy for web or phone

    connect your live agent to your website or phone number so customers can speak with it.

build agents that know what to ask and when to act

build agents that know what to ask and when to act

build agents that know what to ask and when to act

routine requests keep moving with the agent, while the team focuses on the calls that need them.

4 voices, up to 11 languages

choose the voice your customers hear from your agent and let them talk in a language they use every day.

0
voices
siya · ira · aisha · zoya
0
languages

customer context from the first hello

bring customer details from your systems into the call, so the agent gets to the real question without asking for the basics again.

ananya
returning caller
general enquiry
voice agent

“hi ananya, welcome back. how can i help today?”

custom tools for the next step

build custom tools that connect to your systems, so the agent looks up records, creates tickets, or updates information during a call.

calendar
crm
helpdesk
preferred time noted
caller details saved
message sent to team

one agent for web and phone calls

the agent you build for your site can answer a phone line, and your api can place a phone call with that same agent.

web call
how can we help?talk to our receptioniststart call
incoming calloutgoing call
voice agent00:24

a conversation history for every ai receptionist agent

a conversation history for every ai receptionist agent

a conversation history for every ai receptionist agent

look back at what the caller asked and how the agent responded before deciding what comes next

frequently asked questions

frequently asked questions

frequently asked questions

yes, after you add the business information to its instructions in the rumik playground. include the opening hours and service details it should use, along with instructions for questions the supplied information does not cover.
the rumik playground lets you create the ai receptionist agent from a matching template where available or build it from scratch. its instructions define how it handles calls, so setup starts with the business information it needs and how it should respond. once the voice, languages, and any required tools are configured, test a conversation before deploying it.
yes, the same deployed agent can handle both. incoming calls reach it through a connected rented number or sip trunk, while outbound calls are started through rumik's api.
you can instruct it to ask for the caller's message and contact details. to save those details as a separate record or send them to someone, you must configure a custom tool connected to the destination system and add that tool to the agent's instructions.
with silk mulberry 1.6, the ai receptionist agent can speak 11 languages: english, hindi, bengali, gujarati, kannada, malayalam, marathi, punjabi, tamil, telugu, and urdu. its voice settings let you choose which of these languages it should use, along with one of four voices: ira, aisha, siya, or zoya.
its response can be defined in the playground's system instruction field. for example, the instructions can tell it to acknowledge that it cannot confirm an answer and ask for contact details for follow-up. if that request needs to become a support ticket, a custom tool connected to the support system can create it during the call.
on pay as you go, the ai receptionist agent can handle up to 4 simultaneous calls when all account slots are available. each call occupies 1 slot for its duration, and those slots are shared with other agents and text-to-speech requests. unlimited plans offer 1, 2, 4, or 8. capacity above 8 slots is available through sales.
pay-as-you-go pricing is ₹1.99 per minute for mulberry agent calls and ₹2.99 per minute for muga. for a fixed monthly bill, unlimited plans cover agent usage at the selected simultaneous capacity, starting at ₹9,999/month for 1 slot. either option has separate phone-number rental and carrier charges.

build your first voice agent

build your first voice agent