< 1.2s

typical AI response latency

13s

LiveKit p50 dispatch delay (documented)

4 min

LiveKit p90 at 50 concurrent users

2,000+

companies on SignalWire infrastructure

The problem

Job dispatch was designed for batch processing, not phone calls

Callers wait for worker assignment

Every inbound call creates a room, evaluates dispatch rules, queues a job, selects a worker, and initializes an agent process before conversation begins. Ten steps to say hello.

Cold starts punish real callers

LiveKit Cloud shuts down deployed agents after inactivity. Developers report 10 to 20 second cold starts even on paid plans. A caller who waits that long hangs up.

Burst traffic compounds the delay

At 50 concurrent users on the Scale plan, documented p90 latency reached 245 seconds (over four minutes). LiveKit's recommendation: stagger connections with one-second delays.

WebRTC overhead on every call

Each phone call negotiates ICE candidates, STUN discovery, and track subscriptions. These solve NAT traversal for browsers, not for server-side telephony.

Build a Voice AI Agent

from signalwire_agents import AgentBase
from signalwire_agents.core.function_result import SwaigFunctionResult

class SupportAgent(AgentBase):
    def __init__(self):
        super().__init__(name="Support Agent", route="/support")
        self.prompt_add_section("Instructions",
            body="You are a customer support agent. "
                 "Greet the caller and resolve their issue.")
        self.add_language("English", "en-US", "rime.spore:mistv2")

    @AgentBase.tool(name="check_order")
    def check_order(self, order_id: str):
        """Check the status of a customer order.

        Args:
            order_id: The order ID to look up
        """
        return SwaigFunctionResult(f"Order {order_id}: shipped, ETA April 2nd")

agent = SupportAgent()
agent.run()

import { AgentBase, FunctionResult } from '@signalwire/sdk';

const agent = new AgentBase({
  name: 'Support Agent',
  route: '/support',
});

agent.promptAddSection('Instructions',
  'You are a customer support agent. Greet the caller and resolve their issue.');

agent.addLanguage({ name: 'English', code: 'en-US', voice: 'rime.spore:mistv2' });

agent.defineTool({
  name: 'check_order',
  description: 'Check the status of a customer order',
  parameters: {
    type: 'object',
    properties: {
      order_id: { type: 'string', description: 'The order ID to look up' },
    },
    required: ['order_id'],
  },
  handler: (args) => {
    return new FunctionResult(`Order ${args.order_id}: shipped, ETA April 2nd`);
  },
});

agent.run();

package main

import (
	"fmt"

	"github.com/signalwire/signalwire-go/pkg/agent"
	"github.com/signalwire/signalwire-go/pkg/swaig"
)

func main() {
	a := agent.NewAgentBase(
		agent.WithName("Support Agent"),
		agent.WithRoute("/support"),
	)

	a.PromptAddSection("Instructions",
		"You are a customer support agent. Greet the caller and resolve their issue.")

	a.AddLanguage(map[string]any{
		"name": "English", "code": "en-US", "voice": "rime.spore:mistv2",
	})

	a.DefineTool(agent.ToolDefinition{
		Name:        "check_order",
		Description: "Check the status of a customer order",
		Parameters: map[string]any{
			"type": "object",
			"properties": map[string]any{
				"order_id": map[string]any{
					"type": "string", "description": "The order ID to look up",
				},
			},
			"required": []string{"order_id"},
		},
		Handler: func(args map[string]any, rawData map[string]any) *swaig.FunctionResult {
			orderID := args["order_id"]
			return swaig.NewFunctionResult(
				fmt.Sprintf("Order %v: shipped, ETA April 2nd", orderID),
			)
		},
	})

	a.Run()
}

import com.signalwire.sdk.agent.AgentBase;
import com.signalwire.sdk.swaig.FunctionResult;

import java.util.List;
import java.util.Map;

public class SupportAgent {
    public static void main(String[] args) throws Exception {
        var agent = AgentBase.builder()
                .name("Support Agent")
                .route("/support")
                .build();

        agent.promptAddSection("Instructions",
                "You are a customer support agent. "
              + "Greet the caller and resolve their issue.");

        agent.addLanguage("English", "en-US", "rime.spore:mistv2");

        agent.defineTool(
                "check_order",
                "Check the status of a customer order",
                Map.of("type", "object",
                       "properties", Map.of(
                           "order_id", Map.of(
                               "type", "string",
                               "description", "The order ID to look up")),
                       "required", List.of("order_id")),
                (toolArgs, rawData) -> {
                    var orderId = toolArgs.get("order_id");
                    return new FunctionResult(
                        "Order " + orderId + ": shipped, ETA April 2nd");
                }
        );

        agent.run();
    }
}

# frozen_string_literal: true

require 'signalwire'

agent = SignalWire::AgentBase.new(name: 'Support Agent', route: '/support')

agent.prompt_add_section('Instructions',
  'You are a customer support agent. Greet the caller and resolve their issue.')

agent.add_language(name: 'English', code: 'en-US', voice: 'rime.spore:mistv2')

agent.define_tool(
  name:        'check_order',
  description: 'Check the status of a customer order',
  parameters:  {
    'order_id' => { 'type' => 'string', 'description' => 'The order ID to look up' }
  }
) do |args, _raw|
  SignalWire::Swaig::FunctionResult.new(
    "Order #{args['order_id']}: shipped, ETA April 2nd"
  )
end

agent.run

<?php
require 'vendor/autoload.php';

use SignalWire\Agent\AgentBase;
use SignalWire\SWAIG\FunctionResult;

$agent = new AgentBase(['name' => 'Support Agent', 'route' => '/support']);

$agent->promptAddSection('Instructions',
    'You are a customer support agent. Greet the caller and resolve their issue.');

$agent->addLanguage('English', 'en-US', 'rime.spore:mistv2');

$agent->defineTool(
    name: 'check_order',
    description: 'Check the status of a customer order',
    parameters: [
        'order_id' => ['type' => 'string', 'description' => 'The order ID to look up'],
    ],
    handler: function (array $args): FunctionResult {
        return new FunctionResult("Order {$args['order_id']}: shipped, ETA April 2nd");
    }
);

$agent->run();

#!/usr/bin/env perl
use strict;
use warnings;
use lib 'lib';
use SignalWire::Agent::AgentBase;
use SignalWire::SWAIG::FunctionResult;

my $agent = SignalWire::Agent::AgentBase->new(
    name  => 'Support Agent',
    route => '/support',
);

$agent->prompt_add_section('Instructions',
    'You are a customer support agent. Greet the caller and resolve their issue.');

$agent->add_language(name => 'English', code => 'en-US', voice => 'rime.spore:mistv2');

$agent->define_tool(
    name        => 'check_order',
    description => 'Check the status of a customer order',
    parameters  => {
        order_id => { type => 'string', description => 'The order ID to look up' },
    },
    handler => sub {
        my ($args, $raw) = @_;
        return SignalWire::SWAIG::FunctionResult->new(
            response => "Order $args->{order_id}: shipped, ETA April 2nd"
        );
    },
);

$agent->run;

#include <signalwire/agent/agent_base.hpp>

using namespace signalwire;
using json = nlohmann::json;

class SupportAgent : public agent::AgentBase {
public:
    SupportAgent() : AgentBase("Support Agent", "/support") {
        prompt_add_section("Instructions",
            "You are a customer support agent. "
            "Greet the caller and resolve their issue.");

        add_language({"English", "en-US", "rime.spore:mistv2"});

        define_tool({
            .name = "check_order",
            .description = "Check the status of a customer order",
            .parameters = {
                {"order_id", {{"type", "string"},
                              {"description", "The order ID to look up"}}}
            },
            .handler = [](const json& args, const json&) {
                auto order_id = args.value("order_id", "unknown");
                return swaig::FunctionResult(
                    "Order " + order_id + ": shipped, ETA April 2nd");
            }
        });
    }
};

int main() {
    SupportAgent().run();
}

using SignalWire.Agent;
using SignalWire.SWAIG;

var agent = new AgentBase(new AgentOptions { Name = "Support Agent", Route = "/support" });

agent.PromptAddSection("Instructions",
    "You are a customer support agent. Greet the caller and resolve their issue.");

agent.AddLanguage("English", "en-US", "rime.spore:mistv2");

agent.DefineTool("check_order", "Check the status of a customer order",
    new { type = "object", properties = new {
        order_id = new { type = "string", description = "The order ID to look up" }
    }, required = new[] { "order_id" } },
    (args, rawData) =>
    {
        var orderId = args.TryGetValue("order_id", out var id) ? id : "unknown";
        return new FunctionResult($"Order {orderId}: shipped, ETA April 2nd");
    });

agent.Run();

use signalwire::agent::AgentBase;
use signalwire::swaig::FunctionResult;
use serde_json::json;

fn main() {
    let mut agent = AgentBase::builder()
        .name("Support Agent")
        .route("/support")
        .build();

    agent
        .prompt_add_section("Instructions",
            "You are a customer support agent. Greet the caller and resolve their issue.", &[])
        .add_language("English", "en-US", "rime.spore:mistv2");

    agent.define_tool(
        "check_order",
        "Check the status of a customer order",
        json!({"type": "object", "properties": {
            "order_id": {"type": "string", "description": "The order ID to look up"}
        }, "required": ["order_id"]}),
        Box::new(|args, _raw| {
            let order_id = args.get("order_id").and_then(|v| v.as_str()).unwrap_or("unknown");
            FunctionResult::with_response(&format!("Order {order_id}: shipped, ETA April 2nd"))
        }),
    );

    agent.run();
}

Two architectures, one phone call

LiveKit: Job Dispatch

Call arrives, room created
Dispatch rule evaluated
Job created and queued
Worker selected from pool
Worker accepts job
Agent process initialized
Agent joins room
Track subscription established
Audio pipeline ready
Conversation begins

SignalWire: Direct Answer

Call arrives
Routed to the agent
Agent answers
Conversation begins

Documented dispatch latency at 50 concurrent users

Percentile	LiveKit Latency	SignalWire Latency
p50	13,498ms (~13 seconds)	<1200ms
p90	245,331ms (~4 minutes)	<1200ms
p99	276,611ms (~4.6 minutes)	<1200ms

At 50 concurrent users on the Scale plan, a developer documented p90 latency of 245 seconds. LiveKit's response: stagger connections with one-second delays between batches. Source: LiveKit Community Forum.

The solution

Calls as first-class objects, not jobs in a queue

Direct call answering

No job dispatch, no room creation, no track subscription. The agent answers the call. Sub-second response, every time.

One address for everything

Phone numbers, SIP endpoints, AI agents, and subscribers are all addressable the same way. No separate APIs per type.

Native telephony, no bridge

SIP and PSTN handled natively. No WebRTC bridge, no ICE candidates, no STUN discovery, no TURN relay. Media stays on the platform.

Transfers without room migration

Transfer between agents, numbers, and endpoints using standard telephony primitives. No participant migration between rooms.

Architecture comparison

Aspect	LiveKit	SignalWire
Call model	Room with participants	First-class call object
Agent model	Worker dispatched to room	Addressable endpoint
Call answering	Job dispatch (10-50s documented)	Sub-second
Media transport	SIP-to-WebRTC bridge + ICE/STUN/TURN	Native SIP/PSTN
Cold start	10-50 seconds documented	None (HTTP server)
Burst handling	p90 of 245s at 50 concurrent	HTTP endpoint scaling
Transfer	Participant migration between rooms	Standard telephony primitives

Transport is not the bottleneck A published analysis shows the transport layer accounts for less than 5% of total conversational latency. The bottleneck is the model, not the pipe. WebRTC adds complexity without improving voice AI performance.

Get from pip install to production calls

1

Install the SDK

pip install signalwire-agents. One package, one version, zero dependency coordination.

2

Define your agent

Inherit AgentBase, set a prompt, add skills and tools with Python decorators.

3

Point a phone number at it

Configure a SignalWire number to route to your agent. No rooms, no dispatch, no workers.

4

Deploy anywhere

The same code runs on a server, Lambda, Cloud Functions, or Azure. No platform lock-in.

FAQ

Why does LiveKit's dispatch model add so much latency?

LiveKit creates a room, evaluates dispatch rules, queues a job, assigns a worker, and initializes an agent for every call. Each step adds time. Under load, jobs queue behind each other.

Does SignalWire use WebRTC at all?

SignalWire supports WebRTC for browser-based use cases. For phone calls, media flows through the native SIP/PSTN stack with no bridge conversion.

What happens during traffic spikes?

SignalWire agents are HTTP endpoints. They scale the same way any web service scales. No job queue, no worker pool bottleneck.

Can I migrate from LiveKit incrementally?

Yes. SignalWire handles SIP trunking natively. You can route specific numbers to SignalWire agents while keeping other traffic on existing infrastructure.

Trusted by 2,000+ companies

Calls, not jobs. Sub-second, not minutes.

Build voice AI on infrastructure designed for phone calls, not job queues.

Get Started Free Read the Docs

Your Caller Is Waiting in a Queue