AssemblyAI

We just released Universal-3 Pro. And it's free for all of February. LLM-style control over transcription: not just keywords, but actual instructions: → "Be more verbatim, capture filler words" → "This is a medical conversation" → "Label speakers as doctor vs. patient" One prompt. Hundreds of calls. Consistent output.

You probably already know someone who's been talking about adding Voice AI to their product We just launched a referral program to help you help them get started: $100 in AssemblyAI credits for them, up to $100 cash for you Now is the time to share your invite link with that friend you've been brainstorming speech-to-text ideas with Find your custom link in your AssemblyAI dashboard under 'Refer Friends' and send it to whoever comes to mind 👇 assemblyai.com/signup

@kfugere we love to see it 💙🎙️

Everyone is dumping on Granola today. Been a user for a long time but eventually realized I was just copying and pasting the transcripts instead of using their AI capabilities. Started building my own version a couple months back. Saves all notes locally. Bring your own @AssemblyAI and @OpenAI API keys. Costs about $1 per day for 5-6 hours of meetings. fiddleheadai.com

Try it yourself at assemblyai.com/playground

Speaker diarization in real-time streaming just hit a new bar. Check out our demo of speaker labels with Universal-3 Pro Streaming 👇

The best coding agents still need a little help when your API is moving faster than their training data. We've shipped a lot in the last few months: Universal-3 Pro Streaming, Universal-3 Pro, and LLM Gateway to name a few. Features that simply postdate what most coding agents were trained on. So we built the AssemblyAI Skill for Claude Code, Codex, Cursor, and any agent that supports skill/rules files. It gives your coding agent accurate, current knowledge of our full API surface: the right auth patterns, the right model IDs, the right SDK usage, so you spend less time debugging and more time shipping. Drop it into your setup in under two minutes and start building with agents that actually know AssemblyAI. Grab the skill here 👇 github.com/AssemblyAI/ass…

📝 Summarize 0.12.0 out now! Fast summaries from URLs, files, and media - used in @openclaw Big update for slides + media: stronger Chrome sidepanel slide mode, better YouTube/video switching, AssemblyAI+Gemini transcription, xurl for X, NVIDIA provider github.com/steipete/summa…

@steipete @openclaw we love to see it!! 💙🎙️

The era of hardcoded speech pipelines is over. Our team is hosting a deep-dive tomorrow to chat through how Universal-3 Pro and Universal-3 Pro Streaming let you shape model behavior through prompting. No extra post-processing, no brittle workarounds. 🗓️ Tuesday, March 10 📍 San Francisco Topics on the table: 🔹 Entity accuracy — credit cards, phone numbers, emails at sub-300ms latency 🔹 Speaker diarization via prompting — no speaker_labels param required 🔹 Turn detection built for voice agent use cases 🔹 Keyterms vs. prompt — when to use which 🔹 PII redaction and verbatim control in a single prompt Come dig into the real stuff: messy audio, turn detection, entity accuracy, speaker diarization. Bring your burning questions. Register here: luma.com/bv3vo4vj

London builders—our team is coming across the pond! We're hosing an event with @meetgranola: an evening of honest conversation about what it actually takes to build with Voice AI right now. The challenges, the wins, the questions no one has good answers to yet. 🗓️ 25 March, 6:00–9:00pm GMT 📍 London Capacity is limited, and spots are going to people who are genuinely in the weeds building—developers, founders, and product folks working at the intersection of voice and AI. If that's you, sign up to save your spot: luma.com/6agz99u5

AssemblyAI's technical team is hosting a deep dive into Universal-3 Pro and Universal-3 Pro Streaming: what the models can do, how to prompt them, and how to ship faster with less post-processing. Join us on Tuesday, March 10 to check out our newly launched models live and get your questions answered. Topics on the table: 🔹 Entity accuracy — credit cards, phone numbers, emails at sub-300ms latency 🔹 Speaker diarization via prompting — no speaker_labels param required 🔹 Turn detection built for voice agent use cases 🔹 Keyterms vs. prompt — when to use which 🔹 PII redaction and verbatim control in a single prompt Join the team to dig into the real stuff (messy audio, turn detection, entity accuracy, speaker diarization), and open the floor to your burning questions. Register here: luma.com/bv3vo4vj

Try Universal-3 Pro Streaming here: assemblyai.com/universal-3-pr…

We took Universal-3 Pro Streaming out for a spin in the New York subway Spoiler: it's subway-proof 😎

Try Universal-3 Pro Streaming today: assemblyai.com/universal-3-pr…

We tried Universal-3 Pro Streaming in our office with some icebreaker questions. Let's just say it didn't miss a thing.

@perezcarreno Yes!! Code-switching is supported out of the box :)

@AssemblyAI Amazing! Congratulations! Does "Global language coverage" mean it is also multilingual?

Real-time transcription just got a significant upgrade. Universal-3-Pro is now available for streaming — bringing AssemblyAI's most accurate speech model to live audio for the first time. Developers building voice agents, live captioning tools, and real-time analytics pipelines now get three things they've been asking for: 🔹 Best-in-class word error and entity detection across streaming ASR benchmarks 🔹 Real-time speaker labels — know who said what, as it happens 🔹 Superior entity detection for names, places, orgs, and specialized terminology in real-time 🔹 Code-switching and global language coverage built-in

Try Universal-3-Pro Streaming directly on our landing page: assemblyai.com/universal-3-pr…

@Vera_an99 Glad you liked it! 💙🎙️

Stumbled upon AssemblyAI just now. If anyone’s looking for a tool to handle auto-transcriptions or meeting summaries, this looks pretty solid. Just sharing in case it helps someone out! https://www.assemblyai(.)com/

Real-time voice AI just crossed a threshold. It’s no longer a demo. Spoke with @AssemblyAI's @YouveGotFox about promptable speech models and what’s actually changing. Here’s what stood out 👇

@RhysSullivan what are you transcribing? if you can share an audio sample i have a feeling that you can prompt Universal-3 Pro to account for your British accent.. but let me check! -Mart

none of the speech to text tools can get my mut of a british american accent right for example they always interpret me saying `parse` as `pass` or `call` as `cool` - anyone have recommendations?