Podcast Creator Studio

An enterprise-grade studio — for anyone.

Podcast Creator Studio is an enterprise-grade studio that makes anyone the producer of their own podcast or lecture series. It draws from the web and from any information you bring — books, PDFs, texts, notes, or anything else. The best AI models — running on your own API keys for both the writing and the voices, or on ours — then write, voice, and mix full episodes in any language — lectures, interviews, discussions, panel debates, or solo monologues; informational, persuasive, or analytical — performed by characters you design, each with distinct expertise, intellect, emotion and personality. Set a schedule, and a polished episode — verified, fact-checked, and timeframe-relevant — lands in your inbox and publishes itself as a subscribable feed, ready for Apple Podcasts, Spotify, and beyond.

✨ Quick Start — build your first show ▶ Hear a sample first See how it works

Behind the scenes

Three chapters. Zero guesswork.

The exact clicks — no marketing sleight of hand. Cast a character from history, then build a show from either the open web or a book on your shelf.

The Cast Impersonate a historical or fictional figure Turn Marie Curie or Sherlock Holmes into a co-host in about thirty seconds.

01 · CHARACTERS RAIL

Open Characters in the sidebar and click ✨ Impersonate a real person.

02 · NAME THEM

Type any real or fictional name — Marie Curie, Sherlock Holmes, Malcolm X. Wikipedia is looked up in the background.

03 · REVIEW THE DRAFT

The active LLM proposes their expertise, disposition, vocabulary, and emotional register — and offers their Wikipedia portrait as an avatar (opt-in). Edit anything.

04 · SAVE

Your new character is available to any show. Voice cloning stays a separate opt-in — never automatic.

The Wire Build a show from the open web Pick sites, set a freshness window, and generate an episode from today's news.

01 · NEW PODCAST

Click New podcast, name it (Latest Financial News), and pick a Focus so the agent knows what to look for.

02 · SOURCES · SCRAPE URLS

On the Sources tab, keep Source mode on Scrape URLs. Tick source categories (Bloomberg, Reuters, FT…) or paste your own URLs — one per line.

03 · TIME WINDOW

Set Time window to Past 24 hours so nothing stale slips in. Optional: turn on Deep crawl to follow same-domain links.

04 · CAST + GENERATE

Assign your hosts on the Cast tab and press Generate. Fact-checked script → voiced audio → RSS-ready feed, in a few minutes.

III

The Library Adapt a book (or paper, or report) into an episode Drag a PDF, choose "uploads only", and let two hosts discuss it end to end.

01 · NEW PODCAST

Create the show as before — The Fabric of Reality: episode by episode.

02 · SOURCES · UPLOADED FILES ONLY

Switch Source mode to Uploaded files only. The web is now off-limits — nothing outside your library leaks into the script.

03 · DROP THE FILE

Drag a .pdf, .docx or .txt onto the Uploaded materials card. Chunked and indexed in seconds; wait for indexed.

04 · TWO HOSTS · GENERATE

Put two characters on the Cast tab — say, the author and a challenger — and generate. You get a chapter-by-chapter dialogue grounded strictly in the book.

Start free — build your first show No credit card. Bring your own API key or use the studio's built-in AI.

A complete production system — not a script generator

Most tools stop at "an AI wrote something." This runs the whole pipeline — research, writing, casting, voicing, fact-checking, post-production, cost accounting, and publishing — with real depth in every stage.

Research & grounding

📚

Your sources + the open web

websites, RSS feeds, and your own uploads — books, PDFs, notes — are retrieved and grounded into every script; or run knowledge-only / uploads-only when you prefer.

🕸️

Deep relevance crawling

goes far beyond a single page: it walks links to a depth you set, LLM-judges each page against your topic, and ingests only what's actually relevant.

✅

Fact-checked, with source rules

every claim is verified against your sources; unsupported statements are flagged and rewritten — and a minimum-sources rule you control refuses to run on too-thin material.

⏱️

Freshness windows

hour- or day-precise recency windowing keeps each episode to material inside your chosen timeframe, so nothing stale ever slips in.

Writing & cast

🧠

Best-in-class AI, your choice

Claude, GPT, Gemini, DeepSeek, Perplexity, Mistral or local models write in your hosts' voices and your format — with an active→fallback→backup chain so a provider outage never kills a run.

🎭

Designed characters

build hosts with any expertise and dialed-in intellect, emotion, curiosity, and disposition — or base one on a real public figure in a click.

🤝

Built-in AI Assist

an assistant that sets up podcasts and characters for you, answers questions about the app, and can act on your behalf — by typing or by voice.

🌍

Multilingual

auto-translate and re-voice the same episode across many languages, with per-language dialect control.

Voices

🗣️

Lifelike voices

premium text-to-speech with natural crosstalk and overlap, audio tags, and real emotional delivery.

👥

Multi-provider voice cloning

clone a voice on ElevenLabs, XTTS, or F5 — the studio resolves the right engine per speaker and language and automatically switches clone providers when one is unavailable.

🏠

Local & self-hosted voices

run entirely on your own box — free edge_tts, or self-hosted XTTS / F5-TTS / Piper — no cloud and no per-character fees.

🔑

Bring your own keys (BYOK)

your own keys for both the writing AI (OpenAI, Anthropic, DeepSeek, Gemini, Perplexity, Mistral) and the voices (ElevenLabs, Azure, Google, OpenAI) — pay each provider at cost, no markup, no lock-in. Encrypted at rest; or start free on the studio's built-in AI.

Produce, ship & operate

🎚️

Timeline post-production

a two-lane timeline editor — trim, cut, carve, ripple, and layer background music with ducking and fades.

🎬

Video for every episode

each run also renders an MP4 — host avatars, an animated waveform, and chapter cards — generated automatically, ready to upload.

📅

Scheduled & self-publishing

pick a cadence; finished episodes are produced, emailed to you, and published as a subscribable RSS feed ready for Apple Podcasts, Spotify, and beyond.

💰

Real cost tracking

every run logs its true cost — LLM tokens, TTS characters, and any fees — per episode, with provider fallbacks surfaced, so you always know what a show costs.

🗄️

Backups & health

automatic backups of your studio — configs, characters, and data — plus a health panel to monitor and restore.

🔒

Self-hosted & private

your sources, voices, keys, and listener data stay on your own infrastructure — nothing leaves your instance.

Simple, transparent pricing

Start free with your own API key. Upgrade when you want the studio to handle the AI for you.

Free

Bring your own API key. Perfect for trying it out.

✓ 2 podcasts
✓ Your own LLM API key (BYOK)
✓ Free edge-TTS voices
✓ Up to 10-min episodes
— Scheduling
— ElevenLabs / premium voices
— Live Conversation

Start free

Frequently asked questions

The questions people ask before their first episode.

Can I turn a PDF or book into a podcast?

Yes. Upload PDFs, documents, or notes as sources and they're grounded into the script via retrieval, the same way web pages and RSS feeds are — the hosts discuss what's actually in your material, not a generic summary.

Does the AI fact-check what it writes?

Yes. Every claim is checked against your sources; unsupported statements are flagged and rewritten. You control a minimum-sources rule, and a freshness window keeps each episode to material inside your chosen recency range.

Can I use my own OpenAI, Anthropic, or ElevenLabs API key?

Yes — bring your own key (BYOK) for both the AI that writes scripts (OpenAI, Anthropic, DeepSeek, Gemini, Perplexity, Mistral) and the voices that speak them (ElevenLabs). You pay each provider directly at cost, with no markup and no lock-in. Or start free with the studio's built-in AI.

Is my data private? Where does it run?

Podcast Creator Studio is self-hosted — your sources, voices, and listener data stay on your own infrastructure. API keys are encrypted at rest and never leave your instance.

Can it publish episodes in other languages?

Yes. The same episode can be auto-translated and re-voiced across many languages from a single generation run.

Can it publish on a schedule automatically?

Yes. Set a cadence and finished episodes are produced and emailed to you automatically, with an RSS feed your listeners can subscribe to in any podcast app.