Newsletter
Join the Community
Subscribe to our newsletter for the latest news and updates
Open-source toolkit to create custom AI influencer personas with Stable Diffusion, gTTS, and SadTalker.
AI Influencer Generator is an open-source Python project that converts a text prompt into a talking virtual influencer video by chaining Stable Diffusion, gTTS, and SadTalker.
AI Influencer Generator is a free, open-source toolkit for creating virtual AI influencers. It takes text inputs — a prompt to design a persona and a script for the voiceover — and outputs a lip-synced video of that persona speaking. The project runs as a Jupyter notebook, supports Google Colab execution, and is maintained by SamurAIGPT with contributions from Anil Chandra Naidu Matcha and Ankur Singh. The code is released under the MIT License.
The pipeline is: Prompt → Stable Diffusion image → gTTS audio → SadTalker video → AI influencer content. You design a persona, write a script, generate the voice, animate the face, then publish the video. The repository recommends running the provided Jupyter notebook in Google Colab for the quickest setup.
Yes. The core functionality relies on open-source models (Stable Diffusion, gTTS, SadTalker) and requires no paid APIs, although you can optionally use OpenAI GPT for prompt generation.
The tech stack includes Stable Diffusion for image generation, gTTS for text-to-speech, SadTalker for lip-sync animation, and OpenAI GPT for prompt generation.
Not necessarily. The recommended path is to run the notebook in Google Colab, which provides a free GPU runtime. Local setup is also possible, but generating images and videos will be slower without a GPU.
Yes. The project recommends using similar prompts to keep the character's appearance consistent across generated content, which is key for building an influencer persona.
Open the AI_Influencer.ipynb notebook in Google Colab, run each cell sequentially, customize your persona, and download the generated content. For local use, clone the repo and run pip install -r requirements.txt then launch Jupyter.
