Vendor page
Blurt by AssemblyAI: Voice to text, wherever you want it
AssemblyAI, AssemblyAI, 2026
Read the originalRead on October 20, 2026
- What it is
- The official product page of Blurt, a free, open-source macOS dictation app that AssemblyAI publishes. The page says the app runs on AssemblyAI's Dictation API, has no local model to load, and is MIT licensed, and it lists its features and FAQ.
- How much weight it carries
- The vendor's own product page, so it is authoritative for what Blurt is built on and what it offers. Speed and accuracy figures are the vendor's claims. The page does not give a data retention or training statement for audio sent to the API.
Facts taken from this source
- Blurt's product page says you can dictate in 19 languages and switch between them mid-sentence.
“Dictate in 19 languages and code-switch mid-sentence.” (Feature card headed '19 languages')
- Blurt's FAQ says users bring their own AssemblyAI key and pay for what they transcribe, around $0.62 per hour of audio.
“You bring your own AssemblyAI API key and pay for what you transcribe, around $0.62 per hour of audio.” (Is Blurt free? (FAQ))
- Blurt's FAQ says product names, acronyms and function names can be added through custom vocabulary.
“Add your product names, acronyms, and function names through custom vocabulary and they come out spelled right the first time.” (How accurate is Blurt with technical terms and accents? (FAQ))
- Blurt's FAQ says Blurt is free and open source under the MIT license.
“Blurt is free and open source under the MIT license.” (Is Blurt free? (FAQ))
- Blurt's product page says writing styles are defined in a prompt and switched from the main menu.
“One style for Slack, another for email. Define them in a prompt, switch between them from the main menu.” (Your voice, your style)
- Blurt's FAQ says it runs on AssemblyAI's Dictation API.
“It runs on AssemblyAI's Dictation API, the same transcription, punctuation, and formatting pipeline behind production voice products.” (What is Blurt? (FAQ))
- Blurt's product page says the words land wherever the user is typing.
“Hold a key, talk, and the words land wherever you're typing.” (Hero)
- Blurt's product page says there is no local model to load.
“Transcription lands in about 150 ms, cleanup in under a second. No local model to load, no job queue to poll.” (Thoughts to text in milliseconds)