Azeem Tariq

Creator of | AI Engineer | Chatbots, RAG, Voice Agents & Automation

[Azeem Tariq @iaz33m]

Burning through ElevenLabs credits, hitting monthly limits, and paying more?

I built Echova Studio for local AI voiceovers, cloning, custom voices, dubbing, and transcription.
Try it for free. The paid version is 65% off right now.

Note: The entire voiceover of this video is created with EchovaStudio: https://t.co/BXNj9QnmBh

Watch the demo 👇

#EchovaStudio #LocalAI #AIVoice #VoiceCloning #TextToSpeech #ElevenLabsAlternative

[Azeem Tariq @iaz33m]

@SaadPastaDev @MuhammadZohaad @alimemonzx @SaadPastaDev, introduce mcp server in your app.

[Azeem Tariq @iaz33m]

We have added OmniVoice support to Echova Studio.

With this update, Echova Studio now supports AI voiceover creation in 600 languages, opening up a much wider path for creators, educators, marketers, agencies, and product teams working on multilingual content.

Whether you are creating training material, social videos, product explainers, ads, or localized campaigns, you can now build voiceovers for a global audience from one AI voice studio.

Learn more:
https://t.co/WVTQV669BC

#EchovaStudio #OmniVoice #AIVoiceover #TextToSpeech #ContentLocalization

[Azeem Tariq @iaz33m]

@Okikiola_V @FeyiLux Yes

[Azeem Tariq @iaz33m]

Cursor open-sourced Mixture-of-Kittens (MoK), a megakernel designed to keep MoE GPUs working instead of waiting.

On GB300 NVL72: up to 2.37× faster forward performance and 41% higher training throughput—potentially ~29% less GPU time.

Voice-over created with EchovaStudio:
https://t.co/DDTBQAEY2t

#AI #OpenSource #Cursor #LLMTraining

[Cursor @cursor_ai]

We're open-sourcing Mixture-of-Kittens (MoK), our MoE training megakernel for NVL72s.

It fuses all Mixture-of-Experts communication and computation into a single, fully deterministic kernel, and runs up to 2.37x faster than the strongest public baselines.

[Azeem Tariq @iaz33m]

@Okikiola_V @FeyiLux Download free version, its good for most of the poople.

[Azeem Tariq @iaz33m]

@Okikiola_V @FeyiLux Yes its free

[Azeem Tariq @iaz33m]

Still paying every time you generate an image? You don’t have to. 👀

Meet Echova Pixel — a free app that lets you create unlimited images directly on your Mac.

✨ No credits
💳 No subscription
♾️ No generation limits
🔒 Your prompts and images stay on your computer

Create thumbnails, product visuals, characters, artwork, social media content—and much more.
Available now for Apple Silicon Macs with M1 or newer.

Download it free:
https://t.co/3HYYf8cwXj

#EchovaPixel #AIImageGenerator #GenerativeAI #ContentCreator #MacApps #AITools #ImageGeneration #YouTubeCreators #FreeAI #LocalAI

[Azeem Tariq @iaz33m]

@FemiThePundit @FeyiLux checkout
https://t.co/LEsdSELbmH

[Azeem Tariq @iaz33m]

Introducing Echova Pixel ✨

Turn your ideas into images—right on your own Mac.
Echova Pixel is a free image-creation app designed for everyday creators. Describe what you want, generate it locally, and keep your creative work on your computer.

Perfect for:
🎨 Artwork and characters
📹 Video and thumbnail ideas
📦 Product concepts
🍜 Food and lifestyle visuals
💡 Anything you can imagine

Available now for Apple silicon Macs—M1 or newer.
Windows support is coming soon.

Download Echova Pixel for free:
https://t.co/kLNWHroeVm

This is our first release, and I’d love to hear what you create and what you think.
#EchovaPixel #AIArt #ImageGeneration #MacApps #CreativeTools #IndieHacker #ProductLaunch

[Azeem Tariq @iaz33m]

@RohOnChain Amazing

[Azeem Tariq @iaz33m]

@JoelDeTeves @poolsideai Same here, even on basic tasks model failed multiple times and qwen3.6 35b-a3b crushed everything single task.
Laguna is trained to do better on benchmarks but not on real tasks.

[Azeem Tariq @iaz33m]

@MiaAI_lab use temp=1

[Azeem Tariq @iaz33m]

@elvissun @openwhispr Local transcription is a good call when you’re doing audio passes or handling private source files. If you also need TTS, voice cloning, and dubbing around that workflow, EchovaStudio may be useful since it keeps voice/audio work local without cloud upload or API dependency.

[Azeem Tariq @iaz33m]

This is why local voice tooling is becoming more practical now. Fast on-device TTS is great, but the real workflow also needs cloning, dubbing, transcription, and privacy around source audio. EchoVA Studio may be useful for that broader local/offline voice workflow without cloud upload or API dependency.

[Azeem Tariq @iaz33m]

@JonahBlake For a nature-series voiceover, I’d prioritize controllable pacing, a warm narrative voice, and easy revision passes. EchoVA Studio can fit this well if you want to generate and iterate locally instead of sending every take through a cloud voice tool.