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.
TL;DR: Your transcription model might achieve a 5% Word Error Rate, but your meeting summaries can still be completely unreliable if Diarization Error Rate (DER) spikes. DER is the metric that determines whether your system correctly identifies who spoke each word, measured as the sum of three error types: Missed Speech, False Alarm, and Speaker Confusion. For production multi-speaker pipelines, a DER below 15% is the threshold for reliable speaker-labeled analytics; below 10% is the target for clean audio with controlled conditions, such as high-quality meeting assistant output. Accurate speaker attribution directly determines the reliability of downstream LLM summaries and CRM data. Our async pipeline, powered by pyannoteAI's Precision-2 model, delivers up to 3x lower DER than alternatives on conversational speech.
Latency benchmarks for streaming speech-to-text (TTLB and P99)
TL;DR: A voice agent with a 150ms average STT latency sounds fast in a slide deck, but if its P99 spikes to 1.2 seconds, one in every hundred conversational turns breaks. This piece maps the full end-to-end streaming latency budget (network Round Trip Time (RTT), audio buffering, model inference, Voice Activity Detection (VAD) endpointing), explains why P99 and TTLB are the metrics that matter most for production user experience, and shows how to build a reproducible test harness. We cover how our Solaria-1 model delivers first partials under 103ms and final transcripts around 300ms, backed by an open, reproducible benchmark methodology.
Multi-tenant, white-label speech-to-text for platforms
TL;DR: Building a compliant multi-tenant STT layer requires strict data isolation at the key level, granular cost attribution per tenant, and contractual infrastructure guarantees that flow through to your own SLAs. Self-hosting open-source models introduces DevOps overhead, scaling unpredictability, and the absence of built-in tenant isolation features that a managed API provides by default. Managed infrastructure with per-client keys, certified data handling, and all-inclusive pricing removes most of that build cost ($0.20–$0.61/hr with diarization, translation, and entity recognition included, compared with $240K–$480K/yr in dedicated engineering to self-host) but the isolation and attribution architecture still has to be designed correctly regardless of which vendor provides it.
Transforming note-taking for students with AI transcription
Published on Nov 6, 2024
In recent years, fuelled by advancements in LLMs, the numbers of AI note-takers has skyrocketed. These apps are increasingly tailored to meet the unique needs of specific user groups, such as doctors, sales teams and project managers.
Coconote is focusing specifically on students. Leveraging the power of LLMs, their note-taking app can instantly transcribe, summarize, and contextualize notes from lectures.
Read on to find out why they chose Gladia as the primary transcription provider to power their fast-growing project as they expand globally.
About Coconote
Coconote provides an AI-powered note-taking and study assistance app designed for students.
The company focuses on enhancing accessibility to academic resources by offering features like lecture recording, instant transcription, organized notes, summaries, flashcards, and quizzes.
Their primary target audience is college students, with some reach in high school and continuing education.
Available for mobile and web users, the app first launched in the US in April 2024, reaching over 400k downloads in less than 6 months, and is now expanding to other geographies, starting with Japan.
The project's ultimate goal is to democratize access to academic support tools, making high-quality study assistance available to people from varied socioeconomic backgrounds.
Challenge
The Coconote app addresses the challenge of manual note-taking while in class, which has traditionally distracted students from actively listening and learning in lectures.
By automating the capture and organization of lecture content, Coconote allows students to engage more deeply during class without the worry of missing key points or struggling to keep up with fast-paced lectures.
To succeed in this mission, Coconote uses transcription as the backbone of the app. The app transcribes and organizes recorded lectures (in both audio and video formats) into structured notes, summaries, and study aids.
The transcription process involves Gladia as its main provider, handling recordings from diverse academic subjects in multiple languages.
Objectives
To deploy a top-tier transcription API to power the Coconote app, distinguished by the following characteristics:
Ability to effectively capture speech under varying audio conditions, such as echo-laden and noisy lecture halls;
Transcription accuracy and adaptability to specific terminologies across various academic disciplines;
A truly multilingual approach to asynchronous transcription and translation, with core features and add-ons all available in over 100 languages.
Solution
With Gladia, the Coconote team was able to implement:
Instant transcription for lecture recordings and URLs;
Translation to and from 100 languages;
Organized notes per subject matter;
Summaries, flashcards, and quizzes, generated automatically based on the notes.
Impact
By working with the Gladia team to iterate and scale up, they saw a noticeable impact on their rapidly growing user base and retention, with hundreds of happy users reporting remarkable performance of the app’s note-taking capabilities.
4.8 star rating on App Store
420,000+ signups
300,000+ Instagram followers, including celebrity parents like Joe Rogan and Patrick Dempsey
The team at Coconote is just starting to explore the possibilities that Gladia’s transcription brings to their users and is already considering how they can leverage the insights from data to make the app even more tailored to their specific needs.
About Gladia
Gladia provides a speech-to-text and audio intelligence API for building virtual meeting and note-taking apps, call center platforms, and media products, providing transcription, translation, and insights powered by best-in-class ASR, LLMs and GenAI models.
Having read this case study, do you feel like Gladia could be the right fit for your business too?
