BACK TO SELECTED WORK
CASE STUDY 06 / 09Creative / AI
Multi-speaker podcast generation

AgentPodcast-AI

An AI podcast creation app that turns written content into natural multi-speaker conversations using script generation and text-to-speech engines.

PythonGradioGeminiEdge TTSgTTSFFmpeg
AgentPodcast-AI project previewPROJECT / CREATOR
01 / THE WORK

Behind the build.

AgentPodcast-AI lets creators paste articles, research, stories, or notes and convert them into podcast-style audio. Gemini can generate engaging dialogue, while Edge TTS, gTTS, or pyttsx3 handle voice output. The Gradio interface supports one to four speakers, processing progress, and downloadable audio.

02 / ENGINEERING DECISIONS

What makes it work.

01 / CAPABILITYGenerate podcast scripts from articles, lessons, stories, or research
02 / CAPABILITYSupport for solo narration, host conversations, panels, and roundtables
03 / CAPABILITYMultiple TTS engine choices for flexible voice generation
04 / CAPABILITYDownloadable podcast audio through a simple web interface
03 / INTERACTIVE WALKTHROUGH

From input to outcome.

STAGE 1 OF 5

Paste source content or an idea into the app.

04 / ARCHITECTURE

Under the hood.

Multi-Speaker Dialogue Script to Audio Synthesis Pipeline

Audio Diagram 6.1
Source Material

Articles & Research Notes

Script Generation

Gemini Dialogue Writer

Voice Rendering

Edge TTS / gTTS Engine

Audio Output

FFmpeg Merged Podcast MP3

05 / RESULTS & USE CASES

Built to matter.

OUTCOME 01Faster content repurposing into audio
OUTCOME 02Natural multi-speaker podcast experiences
OUTCOME 03Accessible workflow for educators, creators, and researchers
EducatorsContent creatorsResearchersBusiness explainers