Creating a Host
Walk through the host creation wizard to build a new AI host with a name, backstory, voice, and personalized settings.
Getting Started
- Click Hosts & Voices in the sidebar.
- Click Create New Host in the top right corner.
A dialog asks how you want to get started:
- Generate with AI (default): Describe your host and Jellypod creates them for you, then drops you on the new host's page.
- Start from scratch: Build your host step by step in the wizard, including voice clones and personality.
Choose Start from scratch and click Continue to open the wizard described below. The wizard guides you through seven steps (eight if you clone a new voice). You can navigate forward and backward between steps at any time.
Starter, Educator, and Free organizations can keep up to 10 hosts. If you're at the limit, creating another host opens an upgrade dialog instead, pointing to Creator for unlimited hosts. Deleting an existing host also frees up a slot. See Plans and Pricing for details.
Step 1: Name
Enter a name for your host. This appears in scripts, on your podcast website, and everywhere listeners see them.
- Maximum 100 characters
- A character count helper appears below the input as you type
Step 2: Backstory
Write a backstory that defines your host's personality and speaking style.
- Minimum 10 characters, maximum 3,000 characters
- Use the rich text editor to format with bold, italic, and lists
Enhance Backstory
Click the Enhance Backstory button (sparkle icon) in the editor toolbar to have AI expand your description with richer detail. Write at least a few sentences first so the AI has material to work with.
Step 3: Voice Type
Choose how you want to create your host's voice:
- Voice Clone: Clone your own voice with a short audio sample
- Voice Library: Select from the professional AI voice library or use Magic Voice Design
Step 4: Voice
Depending on your choice in Step 3:
- Voice Clone: Record or upload audio samples. See Voice Cloning for details.
- Voice Library: Browse and select a voice from the library, or use Magic Voice Design. See Browsing the Voice Library and Magic Voice Design.
Voice Style (new voice clones only)
When you record a brand new voice clone, an extra step appears right after the Voice step to preview and choose a voice style:
- More Consistent: Stays closest to the recorded voice with the most stable delivery.
- More Expressive: More dynamic and emotive, and supports audio tags like
[whispers].
Listen to both previews, then pick the one that sounds right. This step does not appear for Voice Library hosts or when editing an existing host.
Step 5: Photo (Optional)
Upload an image for your host's profile. Drag and drop or click Select file to browse.
- Accepted formats: JPEG, PNG, GIF, WebP
- Maximum file size: 10 MB
- Square images work best (e.g., 512x512)
Step 6: Reference Materials (Optional)
Attach up to 10 sources that capture how this host writes or speaks. These could be blog posts, essays, podcasts, or video transcripts. Jellypod uses these to match the host's tone and communication style.
- Click the source buttons to add URLs, upload files, or link other content
- A counter displays how many sources you have added (e.g., "Sources (2/10)")
- Click the remove button on any source to delete it
Step 7: Settings
Configure final details under the heading "Almost there...":
- Title: An optional descriptor like "Tech Journalist" or "Industry Expert" (max 100 characters)
- Personality: Describe communication style and personality traits. Click Add Personality to open the editor (the button reads Edit once personality is set). Leave blank to auto-generate from uploaded sources.
A detailed backstory is the single most important factor in creating a realistic host. Include career history, personality traits, expertise areas, and how they communicate.
Frequently Asked Questions
Can I edit a host after creating them?
Yes. You can edit any host at any time from the Hosts page. Click the three-dot menu and select Edit Host to reopen the editor with all current settings pre-filled.
How do I choose the voice model my host uses?
The voice model is set automatically by the voice you pick, so there is no separate model toggle in host settings. When you record a new voice clone, the Voice style step lets you choose between More Consistent and More Expressive previews.
Was this page helpful?