Runtime and provider requirements

Qontent Tube is a local desktop application, but some of the work it coordinates depends on software installed on your computer and services billed by their providers. Check this page before downloading the private preview.

Required for the core voice workflow

Requirement	Why you need it
Windows x64 or Linux x64	These are the platforms included in the current private release. macOS and ARM builds are not included.
A modern browser	The desktop executable opens its private local workspace in your browser. Bun and Node.js are not required.
Internet access	Voice generation and any connected AI or image service call that provider over HTTPS. Authoring, local files, and playback of existing clips remain local.
Your own Deepgram API key	The current release uses Deepgram for text-to-speech. You can write scripts without a key, but you cannot generate new narration until one is configured.

Qontent Tube and La Rebelion Labs are not affiliated with Deepgram and receive no commission for recommending it. We selected it because its free signup credit has been practical for testing: in our own test, generating about two hours of voice used roughly 5% of the available credit at that time. Allowances, prices, models, and availability belong to Deepgram and can change, so check its current terms before relying on that estimate. Create and manage a Deepgram API key.

Your key is saved locally: Windows uses DPAPI for the current user; Linux uses an owner-only secret file. It is not returned to the browser.

Want ElevenLabs, OpenAI, a local model, or another voice provider? Request your preferred voice provider.

Required for finished video rendering

Voice generation does not require these tools. The RENDER workflow does:

• ffmpeg 5 or later, with the libx264 H.264 encoder and AAC audio encoder. Qontent Tube uses it for media conversion, video encoding, narration and music mixing, and final muxing.
• Google Chrome, Chromium, or Microsoft Edge installed on the computer. Qontent Tube drives it headlessly to capture the animated storyboard frames that ffmpeg encodes.
• Enough free disk space for your projects, generated audio, temporary frames, and final renders. Video jobs can use substantially more space than script and voice-only projects.

The app detects both tools and shows a setup checklist. Custom executable paths can be configured when automatic detection fails. See Rendering videos.

Optional AI and image services

These credentials unlock convenience features; they are not required to write scripts, build rosters, generate Deepgram narration, or edit storyboards manually.

AI drafting, storyboards, and translations

Choose one:

• An Anthropic API key for Claude through the Anthropic API.
• An OpenAI API key.
• An OpenAI-compatible endpoint and its required key, including a local server such as Ollama.
• Manual mode, with no AI key stored in Qontent Tube: copy the generated prompt into Claude Code or any other chat tool, then paste the reply back for validation.

A Claude subscription or Claude Code login is not an Anthropic API key and is not used directly by Qontent Tube. Claude Code is usable through the manual copy-and-paste workflow.

Generated scene backgrounds

Automatic background generation needs an OpenAI-compatible image API. By default Qontent Tube uses gpt-image-1; when an OpenAI AI-assist key is already configured, it can reuse that credential. Alternatively, copy the prompt into another image tool and upload the resulting PNG, JPEG, or WebP yourself.

Provider costs and data

The free Qontent Tube preview does not include Deepgram, Anthropic, OpenAI, or other third-party usage. You create those accounts, accept their terms, and pay any charges they impose. Qontent Tube shows provider configuration and usage where available, but third-party pricing and limits can change independently.

Before connecting a provider, review what is sent in AI drafting, Storyboards, and the Privacy Policy.

Quick decision table

What you want to do	What you need
Write and organize Markdown scripts	Supported x64 computer and browser
Generate narration	Deepgram API key and internet access
Draft with built-in AI	Anthropic, OpenAI, or OpenAI-compatible API access
Draft with Claude Code or another chat	Manual mode; no additional key in Qontent Tube
Generate scene backgrounds automatically	OpenAI-compatible image API access
Upload your own backgrounds	No image API key
Render MP4, captions, and chapters	Chromium-family browser plus ffmpeg 5+ with libx264 and AAC