🛠️ AI Tool Lab Daily updates · 500+ AI tools

What Wondercraft Actually Does (From a Podcast-Production Freelancer's Desk)

I run a small podcast-production service for B2B brands that want a show but have no host, no studio, and no editor on payroll. For two years that meant booking voice talent, scheduling recording sessions, and paying an audio engineer per episode - a $400-$800 bill before a single ad dollar came in. Wondercraft changed that part of my business in early 2025, and I have billed it into client work every month since.

Here is the plain version. Wondercraft turns text into a finished, produced podcast. You hand it a script, a blog post, a PDF, or even just a rough idea, and it writes the episode (or polishes yours), narrates it with AI voices, drops in intro and outro music, layers sound effects, generates show notes, and can push the result straight to Spotify and Apple Podcasts. No microphone, no recording session, no engineer.

The honest framing: Wondercraft is a podcast factory, not a recorder. Descript edits audio that real humans spoke; Wondercraft manufactures the audio from words. That makes it the right tool when a client says "we want a podcast" but has nobody to put in front of a mic. It is the wrong tool when you already have two founders recording a real conversation - there, Descript or Riverside is the better buy.

One thing worth naming up front: Wondercraft added an AI agent called Wonda in late 2025 that lets you build episodes by chatting instead of learning an editor. I use it for first drafts, then refine in the timeline. The agent is good for speed; the manual editor is where the quality lives.

---

The Features That Pay the Bills

Script Generation From Anything

You paste a URL, upload a PDF, or drop in bullet points, and Wondercraft drafts a full episode script with host banter built in. For client work this is the time-saver: a 2,000-word whitepaper becomes a 12-minute episode script in minutes. I still edit for voice, but the blank-page problem is gone.

AI Voices and Voice Cloning

The voice library runs 300+ voices across 30 languages on the entry paid plan, climbing to 1,000+ across 50 languages on Pro. Voice cloning lets you train a client's own voice from a short sample - 1 clone on Creator, 5 on Pro. For a personal-branded show, the clone means the "host" sounds like the founder without the founder ever recording. The clones sound close, not perfect, but good enough that casual listeners do not notice.

Music and Sound Effects

This is the part that makes episodes feel produced rather than robotic. Wondercraft ships royalty-free intro/outro music and atmospheric sound beds, and you get full commercial rights to everything you make. For a freelancer selling "a real podcast," the music is what closes the gap between AI narration and something a listener stays for.

Multi-Host Conversations

You can configure two or three AI hosts and generate them talking to each other - back-and-forth, interjections, the rhythm of a real chat. I was skeptical until I heard a two-host episode about cybersecurity; the exchanges felt natural in a way I did not expect from a machine. Used for explainer and roundup formats, it removes the need to source co-hosts.

Show Notes, RSS, and Direct Publishing

Wondercraft generates show notes from the episode and hosts the RSS feed, so you can publish to Spotify and Apple Podcasts without a separate hosting bill. For a client with no podcast infrastructure, that is the difference between "we should start a podcast" and "our show is live on Spotify Friday." The SEO-indexed public share pages also help episodes get found.

Faceless Video Podcasting

A newer layer turns the audio into video: AI avatars, auto captions in 20+ styles, and B-roll generated through models like Veo 2 and Flux. For social clips and YouTube, this means one episode becomes a talking-head video and a dozen short-form posts without a camera. I bill this as a separate add-on because clients always want the video cut.

Credit Model (Read This Before You Price)

Wondercraft runs on credits: one credit equals one minute of generated speech. Free gives 6 credits a month (about 6 minutes of audio) with 40 voices and unlimited script generation. Paid plans: Creator at roughly $25 a month billed yearly (100 credits), Pro at about $45 a month yearly (200 credits), Business from $60 per seat a month, Enterprise custom. Unused credits roll over on paid plans. The credit count is the real constraint when you scope client work - a 20-minute episode costs 20 credits.

---

How to Make Money With Wondercraft

Wondercraft's edge is "a produced podcast with no humans in the booth," so the money lives in services where a client wants a show but has no team. Here is what has worked for me and producers I know.

Path 1: Podcast Production as a Service

This is the core offer. A B2B SaaS or agency wants a podcast for thought leadership but has no host or editor. You build the show in Wondercraft - script from their content, branded AI hosts, music, show notes, RSS, publish - and charge $300-$1,500 per month per show depending on episode count. At three shows at $500 average, that is $1,500 a month against a $45 Pro subscription. The client gets a live, professional podcast; you get recurring revenue with almost no incremental cost per episode.

Path 2: Content Repurposing

Most brands already have a blog, newsletter, or pile of whitepapers nobody reads. You turn one asset into a podcast episode and charge $150-$500 per episode. A 2,000-word post becomes a 12-minute audio file their audience will actually listen to. I sell this as a monthly package - 4 repurposed episodes for $600-$1,500 - and it is pure margin because the generation is minutes of work.

Path 3: Audiobook and Long-Form Narration

Authors and self-publishers want a listenable version of their book without booking a narrator for a week. You feed the manuscript in chunks, assign voices per chapter, and deliver WAV files. Charge $200-$1,000 per book or $50-$150 per finished hour. The voice-clone feature lets an author "narrate" their own book from a short sample - a pitch that closes self-published clients who want their voice without the studio time.

Path 4: Internal Communications Podcasts

Companies want exec updates, training, and company news in podcast form, and internal audiences accept AI hosts far more readily than external ones. You set up a weekly internal show and charge $300-$800 a month per client. This is the steadiest work I have because internal comms never stops and the bar for "good enough" is reasonable.

Path 5: Faceless Video and Social Clips

Take the audio episode and add avatars, captions, and B-roll to make a YouTube video plus short-form cuts for LinkedIn and TikTok. Charge $200-$600 per client per month for the video layer on top of the audio. Course creators and personal-brands are easy targets because they already talk to a camera-shy audience that prefers listening.

The Unit Economics

Blended across those paths at a realistic three to five clients: tool cost $45 a month (Pro, billed yearly). Revenue $1,500-$7,500. Margin sits near 99% because the subscription and a little editing time are your only real costs. The constraint is lead flow and your ear for pacing, not delivery. The credits do the heavy lifting; you do the curation and the client relationship.

One caveat I tell every client up front: the free plan gives 6 minutes a month - it is a demo, not a workflow. And longer episodes show the AI seams, so I cap client episodes around 20-25 minutes unless they accept the tradeoff.

---

What Wondercraft Is Bad At (The Honest List)

1. The free plan is a teaser. You get 6 credits a month - about 6 minutes of audio - with 40 voices and unlimited script writing. That is enough to test the workflow, not to run a show. Paid starts at roughly $25 a month (Creator, annual), which gives 100 credits. If a client needs more than 100 minutes a month, you are on Pro at $45, and a weekly long-form show pushes toward Business.

2. Longer episodes expose the machine. Reviewers and my own ears agree: under 15 minutes, most casual listeners do not catch that the hosts are AI. Past 30 minutes, the rhythm flattens and discerning listeners notice. For a daily news brief this is fine; for a 90-minute deep-dive interview, it is not the tool. I keep client episodes in the 12-25 minute band for this reason.

3. Emotional control has a ceiling. Director Mode lets you steer tone line by line, but you cannot say "make this part more excited" with the granularity a human director would. The inflection is good, not great. When a client wants a specific emotional arc, I generate, then lean on the timeline editor for manual pacing - the AI gets me 80% there.

4. Voice fidelity trails the top-tier stack. Against ElevenLabs plus a Riverside recording, Wondercraft's narration is a step behind on pure sonic quality. It is a podcast-specific workflow, not a voice-lab. For high-end audiobook work where the voice is the product, I would still reach for ElevenLabs. For content-marketing podcasts where workflow speed matters, Wondercraft wins.

5. Support is slow and the tiers constrain. Users report slow support responses; the community forum often answers faster than the team. And the feature ceiling bites: several advanced controls sit on the higher plans, so a serious producer outgrows Creator quickly. Budget for Pro or Business if podcast work is your actual business, not a hobby.

6. It manufactures, it does not replace real conversation. Wondercraft shines when no human host is available. When a client has two founders who love recording together, forcing their show through AI voices loses the thing that made it good. Use Descript or Riverside for real-human audio; use Wondercraft for the no-host shows.

---

Wondercraft vs the Alternatives

The summary: if you have real humans talking, Descript or Riverside. If you want the highest voice quality money can buy, ElevenLabs plus a recorder. If you want a complete podcast - script, hosts, music, publishing - built from a document with nobody in a booth, Wondercraft is the pick. For a freelancer selling "you have a podcast by Friday," that end-to-end workflow is the whole pitch.

---

Getting Started Without Wasting Money

• Treat the free plan as a test drive, not a workflow. Six minutes a month lets you hear the voices and generate a script, not ship a show. Run two or three sample episodes to learn which voices match your clients before you pay. Once you know your go-to host voices, the subscription earns back on the first paid job.

• Start on Pro, not Creator, if podcasting is your business. Creator at ~$25 gives 100 credits and 1 voice clone - fine for one casual show. Pro at ~$45 doubles credits to 200, adds 5 voice clones, AI sound-effect generation, and high-fidelity WAV export. If you serve paying clients, the clone count and credit ceiling on Creator will pinch within a month. Pay for Pro and protect your margin.

• Clone the client's voice early, with consent. Voice cloning is the feature that sells personal-branded shows - a founder "hosting" without recording. Get a short clean sample and train the clone on Pro. Always disclose the AI use to the audience; the trust cost of hiding it is higher than the novelty gain.

• Cap episode length for external audiences. Keep client episodes around 12-25 minutes. Under 15, listeners rarely notice the AI; past 30, they do. If a brief demands a long-form deep dive, record real humans or set expectations that the format is synthetic.

• Sell the workflow, not the voices. The line that closes clients is "you get a published podcast on Spotify by Friday, no studio, no host." Lead with the end-to-end delivery and the commercial rights, mention the AI as the how. Buyers fear the production lift more than they care about the tech.

• Add the video layer as a paid add-on. Once the audio episode exists, the faceless video version - avatars, captions, B-roll - is a separate $200-$600 a month of revenue for almost no extra generation cost. Course creators and personal brands take it every time. Bundle it; do not give it away.

---

The Bottom Line

Wondercraft is the tool I open when a client wants a podcast but has no host, no studio, and no editor. It does one thing the recorders do not: it manufactures a produced show - script, voices, music, show notes, RSS, publish - from a document, with nobody in a booth. The free plan is a teaser, longer episodes show the seams, and the voice fidelity trails a top-tier stack - I do not pretend those away.

The money is real and I live it: a ~$45 monthly Pro sub, clients paying $300-$1,500 a month for produced shows and $150-$500 per repurposed episode, three to five clients clearing $1,500-$7,500 against almost no other cost. The work is curation, pacing, and client management, not engineering. If you sell content services and you are tired of the recording-and-editing tax, Wondercraft is the quiet edge that pays for itself on the first retainer.