OCDevel
WalkPodcast

An AI podcast on whatever you want.

Describe a show once. OCDevel Podcaster researches, writes, and voices every episode on its own - and drops them into your podcast app via RSS. Or upload your own audio and let it handle the midroll injection.

Free forever with ads ·

Kokoro · 50+ voices across 9 languages
Qwen3-TTS · higher quality and voice cloning

How it works

1

Describe your show - topic, segments, length, voice.

2

Podcaster researches, writes, and voices each episode on your schedule.

3

Listen in any podcast app via RSS.

Everything Podcaster does

Tap any capability to jump to the details.

AI episode generation

Describe a show once. Podcaster researches, writes, voices, and publishes every episode on a schedule.

Learn more

AI episode generation

Auto-mode, powered by Gemini. Lay out your show as a set of segments and Podcaster runs the whole production loop on its own - no script, no recording, no editing.

Deep research, segment by segment. each segment has its own prompt, target length, and category. Flag one for live internet grounding and it pulls fresh sources - category-aware, so a Technology segment researches differently than Comedy.

Stitched into one cohesive episode. a synthesis pass weaves the researched segments into a single naturally-flowing narration, then writes the title, teaser, and show notes for you.

On your schedule. daily, weekly, biweekly - new episodes generate and land in subscriber apps automatically.

Steerable. drop a feedback note and the next run adjusts; seed specific topics into the queue, or let it refill itself when it runs low.


Bring your own agent

Want full control of the writing? Point your own AI agent at the Podcaster API and it owns the research and prose - while Podcaster still handles voicing, publishing, and the queue.

Full show context on tap. your agent fetches a skill prompt with recent episodes, the pending queue, your segments, and your steering notes - then researches, writes, and submits the episode.

Your tokens, not ours. the agent runs on your LLM budget, so generation scales without limit. OCDevel only charges for the TTS synthesis.

The server keeps the queue healthy. it tells your agent exactly when to seed new topics, so the feed never runs dry and never double-writes.

Clone and run. a ready-made setup repo - drop in an API key and go.

See the agent setup repo →

50+ voices & cloning

Pick from a deep library of natural voices, or make the show sound like you. Preview anything before you commit.

50+ voices, 9 languages. English, Spanish, French, Italian, Hindi, Japanese, Mandarin, Portuguese and more, powered by Kokoro - fast and free forever.

Clone your voice. record a short clip in the browser or upload one (no transcript needed) and Podcaster clones it with Qwen3-TTS for about $0.85/hr of audio.

Design & default. design a voice from a text description, set a per-show or per-episode default, and publish voices to a shared library.

Custom pronunciation. teach it your jargon, acronyms, and code so it never mangles the hard words.


Smart midroll injection

Upload your own audio (or use any generated episode) and Podcaster places pre-, mid-, and post-roll where a break feels natural instead of jarring.

Finds the real break points. it transcribes the audio and scores topic transitions, silence gaps, and sentence boundaries to drop the ad at a natural seam.

Context replay. after a midroll it replays a few seconds of what came before, so listeners pick the thread back up.

Two timing modes. fixed interval or evenly dispersed across the episode.

Swap an ad, re-roll in bulk. change a roll and affected episodes are flagged stale so you can re-inject without regenerating the audio.


Manual episodes & humanize

Not everything needs the AI loop. Bring your own content in whatever form it already exists and Podcaster turns it into a clean episode.

Any source. paste text or markdown, upload a file (.md/.txt, plus EPUB/PDF on paid), or pull from a URL or Google Doc.

Humanize. turn it on and an LLM rewrites code blocks, lists, and tables into prose that actually sounds good read aloud.

Audio-only path. upload a finished recording and use Podcaster purely for midroll injection.

Reprocess anytime. swap the voice or re-run a single stage, with version history so you can always restore.


Publish everywhere

Podcaster handles the plumbing so your show works like any other podcast, on every platform.

Real RSS. subscribe in Apple Podcasts, Spotify, Overcast, or any podcast app via the auto-generated feed.

Cover art & metadata. AI-generated cover and description (paid), iTunes categories, a custom slug, and a public show page.

Your call on visibility. keep it unlisted (feed works, just not in the directory) or list it in the public directory.

Text feed for e-readers. pull a text-only variant of the feed to read instead of listen.

Which audio tool is right for the job?

Podcaster is built for autonomously generating an AI podcast feed - describe a show once and it researches, writes, and voices every episode on a schedule, with 50+ voices plus cloning, RSS, smart midroll injection, and the option to bring your own agent. It's made for learners, hobbyists, and solo podcasters who want a show on autopilot. For other audio jobs, here are the tools I actually recommend:

ElevenLabs

Reach for ElevenLabs when quality is non-negotiable - the industry-benchmark voices everyone else gets compared to, with emotion control, dubbing into 90+ languages, sound effects, and conversational AI in one polished, client-ready studio. Pro voice cloning from $11/mo.

Explore ElevenLabs →

Fish Audio

Reach for Fish Audio when you are shipping at scale or on a budget - near-ElevenLabs quality at roughly 6x cheaper on the API ($15 per 1M characters), 10-second voice cloning, and an open-weights model you can self-host. The developer pick when cost and control beat brand polish.

Explore Fish Audio →

Audible

Reach for Audible when you would rather listen than generate - over a million professionally human-narrated books (real narrators, not AI), Whispersync to pick up where your Kindle left off, and the best app for offline listening. The free trial includes credits to start your library.

Try Audible Free →

Start your podcast in minutes.

Free forever with ads. $10 unlocks cloning, midroll, and bring-your-own-agent - credits never expire, no subscription.