AssemblyAI · 2025
7 tracked updates · 2025
Across 2025, AssemblyAI shipped 7 tracked updates. That's about 1 a month across 7 active months. The busiest month was January 2025 (1). The dominant theme was feature updates (1). The standout: “Introducing new products and model updates to help you build, deploy, and scale Voice AI applications” (Oct 22, 2025).
Browse other years from AssemblyAI
October 2025
AssemblyAI introduced three new products—Speech Understanding, Guardrails, and LLM Gateway—alongside model improvements like Universal-2 and Slam to streamline Voice AI application development. These …
September 2025
AssemblyAI introduced Keyterms Prompting for its Universal-Streaming speech-to-text service, allowing users to boost transcription accuracy for up to 100 custom terms in real-time. The feature improve…
August 2025
AssemblyAI’s Universal model now supports 99 languages at a flat $0.27/hour rate, including automatic language detection for all and speaker diarization for 95 languages. The update eliminates languag…
April 2025
AssemblyAI has launched Slam-1, a prompt-based Speech Language Model now in public beta, representing a fundamental shift in speech recognition by combining large language model reasoning with special…
March 2025
AssemblyAI has been named to Fast Company’s 2025 list of Most Innovative Companies, validating its leadership in Speech AI innovation alongside industry giants like Nvidia, Waymo, and Duolingo. The re…
February 2025
Google Cloud’s AI trends report highlights AssemblyAI CEO Dylan Fox’s insights on ‘last mile’ challenges in AI adoption for startups, emphasizing the need to address industry-specific gaps like real-t…
January 2025
Synthesia, an AI-first video editing platform founded in 2017, shifted its initial vision from building tools for the video industry to creating a collaborative platform that makes professional AI-gen…
Track AssemblyAI on autopilot