from right. you, past the results first quarter. welcome, operator, our everyone, and review to All operating Thank I'll our earnings call.
company. new a This This quarter, use is our management. disciplined we growth achieved and cost success for a GAAP of milestone cases strong profitability, result significant revenue from
had macroeconomic work, from changes the profitability. challenges dynamic faced years, pressures, innovation brought to team's to and over we our While back hard the and focus few regulatory past
sustainable remain ahead and to driving growth. XXXX, As profitable to efficiency committed we look we operational delivering
fourth the quarter up in revenue were year-over-year. million, X% $XXX Total
low-margin live driven such sales, mainly Excluding business and certain in cases shopping of products, by as use expansion certain entertainment. end revenues from
Shengwang, X,XXX and increase excluding respectively, for to the ago. end over X,XXX customers and active X%, to had of close of As X% we for Agora compared Easemob, year of XXXX, X an
our The generative AI and Now updates turn an product for to technology let's business revolution advancing pace. incredible is the quarter.
past well Gemini open the like presents few remarkable oX we've source groundbreaking as DeepSeek. Multimodal our X.X to model extraordinary developments from Google's as This business. Over innovations OpenAI's for months, opportunity seen transformation recent
positioned AI Our beyond and models. realtime ensure is expressive basic conversations, seamless technology engagement and enable chat and multimodal human key technology to enable interactions uniquely By web-based more moving between are intuitive experiences. text the to communication we to
use is new efficient. text not simply or entirely where introduced cases improved also practical It
today many interaction large still capabilities. However, models language voice lack robust
and for text-based easily can developers conversational deliver many developers models, agents. language low-latency conversational language modules, building large to understanding generation, rapidly to this large AI enable speech Essentially, create model any models, speech blocks to argument other and latency, integrates enables open interaction, supported sensitivity is the those Transformative can Extensions lead to TEN For we often process To that short operation to that address to framework support our experience. or community source and this do Network an with falls enabling multimodal build experience. It a AI gap, background framework high to text realtime in experiences. project noises, upward to experience user voice which TEN. the video frustrating an retrievable developer text, interruptions contributed useful useful caused voice-based
For text-to-speech language large STT, agent audios combining example, quickly fish matter model of minutes. DeepSeek build Microsoft developer this can in and a to a all use
of Earning become fastest-growing the has over stars Since months. X,XXX project its TEN its a just GitHub in launch, kind. few
-- the on GitHub, several the the as days. and remained topped multiple list of TEN occasions, for #X On the on project on list list trending patent
conversational empowers accessed that create our apart conversational I'm any build on success What leading is new interactive of to experiences excited AI API, integrate deployed or model. announce voice to it's customer models. Building humans. its enterprise-grade the model Whether model ability solution via engine, a can to launch developers solution naturally TEN's talk an language with large easily our engine AI make it with this interact set AI on-premise and with AI
with integration Multimodal in instance, integrated For same in of for Gemini's The API with within DeepSeek hours December. Google XX January. live true release last XX-hour its was
advantage. solution and also TDS broad is not is STT with many modules. popular compatible Our But compatibility only
engine AI X key conversational elevates experience areas. in Our the conversation
to conversation intelligently intuitive. ensuring engine feel tax and dynamics, milliseconds interactions and XXX for smooth human the create introduction, First, conversations within life process natural user
voice solely our speaker the -- topic to the in from filtering speakers background. irrelevant primary and speaker prime algorithm attention disruptions enables Second, AI focus models selective processing, on the out
our Additionally, for cancellation inputs high-quality AI crystal echo ensure model human the and [indiscernible] that clear generating noise advanced receives suppression responses. necessarily
different in Leveraging queries to infrastructure, Lastly, network latency. minimize our little and packet loss engine global as across milliseconds, response dynamically the conditions. delays XXX handling as
test assistance will to been a AI support. AI conversational across of of We use language cases become adoption with and accelerate customer virtual IoT conversational the smart this and key various virtual have future driver working build devices, across believe closely shopping We to from our industries customers live growth. companions and a applications tutoring innovation
Agora set to at I to up, wrap a I want incredible Shengwang. take and teams Before moment
nothing human interactions and pushing dedication are for work Our we have environment in possibilities of inspiring. unlocking boundary fast-evolving AI and been the communication. dedication Today, of new realtime -- short hard your this Europe AI
and more what lies sharing forward soon. you ahead, look I'm excited about to with updates I
let that, will things financial turn With over to who Jingbo, me review our results.
