Dear Reader,

I spent the majority of time this and last week keeping my poorly kids entertained at home while trying to look for contract work and create an online course, with video. It became clear very quickly that I wasn't going to be able to do everything, but I had nobody else on hand to help me hit my targets. That is, unless you count...me.

On one of those days I had an unexpected two-hour gap where there was nobody in the house - that's all the time it took for me to create a new AI clone of myself to help get some of my heavy lifting done.

As we move deeper into 2025, AI continues to transform how we create and consume content. This week, I’m diving into the latest advancements in voice cloning and text-to-video technology that are making professional content creation more accessible than ever.

Let's get to it.

Text-to-Video: From Words to Visual Stories

Text-to-video technology has evolved dramatically in recent years, with several key players transforming how we create video content. Synthesia, which bills itself as “the world’s first AI video communications platform,” launched Synthesia 2.0 in December 2024, allowing users to transform texts, PowerPoints, PDFs, or URLs into professional videos.

While Synthesia offers impressive capabilities like multilingual translation across 120+ languages with automatic updates it’s important to note that it’s primarily designed for corporate-style videos rather than storytelling or cinematic visuals. The platform excels in training videos, internal communications, and customer support content.

Other significant players in this space include:

• ​Runway: A powerful AI video creation platform with features like text-to-video, image-to-video, and advanced camera controls, although it's more creative in nature, akin to Sora or MidJourney.
• ​HeyGen: Another leading competitor offering AI avatar technology and video generation capabilities. HeyGen is very popular among the AI influencer crowd as it allows users to easily chunk up videos in formats ready for social media.

For businesses and educators, these text-to-video platforms offer several advantages:

• Elimination of expensive studio time and equipment
• Access to pre-designed templates (over 300 in Synthesia’s case)
• Automatic captioning in multiple languages
• Royalty-free media libraries

Check the results from Synthesia below:

📹 If you cannot see the video above, ​​please click here​​.

This was my second take with creating a Synthesia clone. The first time, I didn't take into account my background, lighting, how I was framed in the shot, or where my eyes were looking! This second time around, I'm much happier with the results. Synthesia does take about 24 hours to turn a clone around, so it's worth trying to get these things right the first time.

The future of this technology looks promising, with Synthesia planning to introduce interactive capabilities like clickable hotspots, embedded forms, quizzes, and personalized calls-to-action.

There are some limitations to the platform though, which I'm not completely sold on yet. I'm only limited to 10 minutes of video creation a month on the lowest plan. I was able to "shoot" 7 course intro videos and keep them all around a minute, but I'm not at the point where I'm generating full courses with it...yet.

However, potential users should evaluate which platform best suits their specific needs, as each has different strengths in terms of avatar realism, customization options, and output quality.

Voice Cloning: Finding Your Digital Voice

Voice cloning technology has also made remarkable strides, with the market projected to grow annually at a 26.1% CAGR from 2023 to 2030. Today’s tools can create incredibly realistic synthetic voices from just short audio samples.

Several platforms stand out in 2025, with ElevenLabs continuing to lead innovation in the space. The company just made headlines this week by expanding beyond voice generation with the launch of Scribe, its first standalone speech-to-text model. This new offering claims an impressive 97% accuracy rate for English, outperforming competitors like Google Gemini 2.0 Flash and OpenAI’s Whisper Large V3 in benchmark tests.

What makes Scribe particularly impressive is its support for over 99 languages, with 25 languages achieving “excellent accuracy” (less than 5% word error rate), including English, French, German, Hindi, Japanese, Polish, Spanish, and Ukrainian. The model includes advanced features like speaker diarization, word-level timestamps, automatic sound event tagging, and direct video transcript capabilities.

Other notable voice AI platforms include:

• ​Resemble AI, focusing on security and deepfake detection alongside professional voice cloning
• ​Play.ht, providing high-fidelity cloning with extensive voice control settings

The most exciting developments include multilingual capabilities that maintain the speaker’s unique vocal characteristics across languages, and emotional expressiveness that conveys subtle nuances from joy to empathy.

I've been using ElevenLabs for more than a year, and run my podcast on it, using a mix of the system voices that are available, as well as a re-recorded clone of my own voice that I "remastered" this week.

Your AI and Tech Brief for F... Feb 27 · The Syntellicast 6:12   ​

I haven't found anything better than ElevenLabs yet - for my money it's the one to watch.

---

​

Made with ❤️ in Valencia by Jim Christian. For feedback, please reach out to hello@jimchristian.net.

​