AI Podcast Generator
Explore the AI podcast generator
You can try every option before you buy. Generating audio needs a paid plan.
Sign in or sign up and choose a plan to generate your first podcast.
Watch the video tutorial
Paste a text, upload a file or import a web page. Choose a template, the models and two voices, and the generator turns your content into an AI podcast with two hosts.
How to make an AI podcast
Seven steps take you from a text to a finished podcast with two hosts.
1. Add your content
Paste a text, upload a PDF or TXT file, or import a web page. The text model writes the dialogue from this content.
2. Choose a template
A template sets the style and the language of the dialogue, for example a podcast, a debate or a lecture. To have your own script read as written, choose “read directly from a script” and write one line for each turn:
speaker-1: Welcome back to the show. Today we look at how bees talk to each other. speaker-2: Mostly with a dance. It sounds strange, but it works.
3. Choose the models
The text model writes the dialogue, and the voice model reads it:
- Flash v2.5 (fast): The lowest cost, at 0.5 credits per character. Good for drafts and long podcasts.
- Multilingual v2: Natural, steady speech in many languages, at 1 credit per character.
- Eleven v3 (expressive): The widest emotional range, with audio tags such as [laughs] or [whispers], at 1 credit per character.
4. Choose two voices
Choose a voice for each host. The play button next to the list plays a sample of the voice.
5. Choose the length
A longer podcast uses more credits. Your plan sets the longest podcast.
6. Generate the podcast
Select “Generate the podcast”. It takes a few minutes. The audio appears on this page when it is ready. If you leave the page, the podcast still appears in My AI Podcasts.
7. Edit and generate again
The transcript goes back into the text box as a script. Change a line, a name or a number, and generate the podcast again.
Plans and limits
| Feature | Hobby | Freelancer | Professional | Enterprise |
|---|---|---|---|---|
| Voice models | Flash v2.5 (fast) | Flash v2.5 (fast), Multilingual v2 | Flash v2.5 (fast), Multilingual v2, Eleven v3 (expressive) | Flash v2.5 (fast), Multilingual v2, Eleven v3 (expressive) |
| Text models | gpt-5.6-terra, gpt-5.6-luna | gpt-5.6-terra, gpt-5.6-luna, gpt-5.6-sol | gpt-5.6-terra, gpt-5.6-luna, gpt-5.6-sol | gpt-5.6-terra, gpt-5.6-luna, gpt-5.6-sol |
| Longest podcast | 1,000 words | 2,000 words | 2,800 words | 3,500 words |
| Longest source text | 3,000 characters | 8,000 characters | 15,000 characters | 40,000 characters |
| Longest script | 10,000 characters | 20,000 characters | 30,000 characters | 40,000 characters |
| Credits per 30 days | 10,000 | 25,000 | 60,000 | 150,000 |
| Voices | 18 | 42 | 110 | 110 |
| File and web page import | No | Yes | Yes | Yes |
| Show notes | No | Yes | Yes | Yes |
| Cover images | No | 25 cover images per 30 days | 60 cover images per 30 days | 150 cover images per 30 days |
| Podcast assistant | No | Yes | Yes | Yes |
Every plan can start 6 podcasts per hour, one at a time.
Tips for a better podcast
- Try two voices with a short script first, to hear how they sound together.
- Punctuation shapes the speech: a comma makes a short pause, and a question mark changes the tone.
- Write names, numbers and abbreviations the way a host says them, for example “twenty twenty-six” for 2026.
- For a long source, make one episode for each chapter or section.
Frequently asked questions
What is AIdeaFlow Podcast?
An AI podcast generator. A text model turns your content into a dialogue between two hosts, and ElevenLabs voices read the dialogue as audio.
Do I need a paid plan?
Yes, generating audio needs a paid plan: Hobby, Freelancer, Professional or Enterprise. Without a plan, you can look at every option. Your plan sets the models, the voices, the longest podcast and the credits, and every plan can start 6 podcasts per hour.
How do credits work?
Each character that the voices read uses credits: 0.5 with Flash v2.5, and 1 with Multilingual v2 and Eleven v3. The credit limit of your plan applies to the last 30 days. A failed generation uses no credits.
Which models can I use?
OpenAI text models write the dialogue. The ElevenLabs models Flash v2.5, Multilingual v2 and Eleven v3 read it. Your plan sets which models you can choose.
How many voices can I choose from?
The Hobby plan has 18 voices, the Freelancer plan has 42, and the Professional and Enterprise plans have all 110.
Can I use my own script?
Yes. Choose the template “read directly from a script” and start each turn with speaker-1: or speaker-2:. A line without a label continues the turn above it. Text before the first label, or a script without any label, is read by host 1.
Which languages can my podcast use?
The templates write dialogues in 18 languages, and the voice models speak many more. With your own script, you can use any language that the voice model speaks.
Can I import files and web pages?
Yes, on the Freelancer, Professional and Enterprise plans. Upload a PDF or TXT file, or import the text of a web page.
Who can listen to my podcasts?
A public podcast appears in the featured podcasts, and anyone can open its page. A podcast that is not public is visible only to you, under My AI Podcasts when you are signed in.