Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.
Building real-time multilingual ASR with code-switching
When a speaker switches languages, traditional models keep outputting the previous one for several hundred milliseconds before catching up, producing garbled text and inaccurate timestamps. The obvious fix is a large multilingual model. But those are expensive to run, awkward to deploy on-device, and still stumble on fast switches.
Factors affecting the accuracy of speech-to-text transcripts
TL;DR: Production STT accuracy fails not because of model benchmarks, but because of the gap between studio evaluation audio and the messy, multilingual, overlapping speech real users produce. Four root causes drive that gap: input audio quality, speaker traits (accents, code-switching, and overlap), domain vocabulary deficits, and model training data diversity. WER alone doesn't capture production risk. Semantic accuracy and Diarization Error Rate matter just as much when CRM syncs, coaching scores, and AI summaries all depend on what the transcript gets right. Solaria-1 delivers on average 29% lower WER on conversational speech and 3x lower DER compared to alternatives, benchmarked across 7 datasets and 74+ hours of audio with open, reproducible methodology.
Business call transcript analysis techniques for sales and support teams
TL;DR: Upstream transcription errors compound through every downstream system: LLMs, sentiment models, and CRM pipelines are only as reliable as the transcript they process. Core conversation intelligence techniques, including sentiment scoring, BANT extraction, objection mining, and talk-ratio analysis, all depend on transcription quality. Async/batch processing provides full conversation context, making it the right default for post-call workflows.
How to build a voice-to-text Discord bot with Gladia real-time transcription API
Published on Sep 21, 2023
Discord, the leading communication platform for gamers and communities, is designed for seamless communication with other users, be it through text channels, DMs, 1-1 calls or even collective voice channels.
Based on multiple request from our Discord members, we’ve built a custom JavaScript bot that makes use of Gladia’s live transcription API to transcribe speech in real time directly on the Discord server.
What can you do with Discord bot?
First, you can transcribe voice in real time directly on Discord’s voice channels. Ex. you’re streaming a game on Discord and want to access some learnings and tips received during the sessions. Or, you’re having your group gathers on the platform and want to be able to review the talking points after – just like with any other virtual meeting platform.
Beyond that, a bot like this could be used for real-time moderation to flag hate speech and ban users. With additional tools like ChatGPT, you could also create command-based notes to provide meeting summaries and helps you catch up with meetings you may have missed.
How to implement the Discord.js v14 bot + Gladia real-time transcription
Step 1: Register your bot
Create a Discord bot that you'd like to use for transcription. If you’ve never built one before, here’s a useful resource to help.
First, install all the required package by running:
npm install
Then, you will to setup the index.js script with your Discord keys, guild ID (Server ID), and the Voice Channel ID.
Step 2: Retrieve API key
Sign up for our speech-to-text API at app.gladia.io and obtain your API key. Documentation for Gladia live transcription can be found here.
Step 3: Code integration
Once everything is set up properly, simply run:
npm run start YOUR_GLADIA_TOKEN
Your bot should then join the channel corresponding to the channel ID you configured in the index.js file.
Step 4: Configure Discord permissions
Make sure your bot is invited on the server;
Give the bot the required voice permissions.
Bear in mind that the current v1 implementation of the bot is not fully optimized, so you might experiences inaccuracy regarding language changes & words.
We hope you enjoyed this short tutorial. Given how much audio data still goes to wasted, we’re always curious to explore the many ways in which transcription tech can be used to remedy that. Let us know if you went on to build a bot or used our API for others apps on Discord or beyond, we’d love to hear from you.
About Gladia
At Gladia, we built an optimized version of Whisper in the form of an API, adapted to real-life use cases and distinguished by exceptional accuracy, speed, extended multilingual capabilities and state-of-the-art features, including speaker diarization and word-level timestamps.
Contact us
Your request has been registered
A problem occurred while submitting the form.
Read more
Speech-To-Text
Building real-time multilingual ASR with code-switching
Speech-To-Text
Factors affecting the accuracy of speech-to-text transcripts
Speech-To-Text
Business call transcript analysis techniques for sales and support teams
From audio to knowledge
Subscribe to receive latest news, product updates and curated AI content.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
.png)
