Transparency

Models & voices

Every AI call the studio makes, whose key it runs on, and what it costs you. The short version: your key, provider list price, $0 markup.

The stack

What runs where

Four kinds of AI touch your reel. Here is each one, plainly.

Script models

Scripts are written by Gemini or OpenAI GPT models running on your own key, grounded in your channel and your knowledge base so nothing drifts off-topic.

Prefer Claude? Through the built-in MCP server, Claude Desktop, Claude Code, and Cursor can drive the whole studio — drafting, scenes, rendering, publishing — under your review.

Image models

AI scene stills come from Gemini / Imagen or OpenAI image models, again on your key at the provider's per-image price.

Images are optional. Most scenes are motion-graphic compositions — code cards, terminals, charts, diagrams — which cost nothing to generate.

Voice

Narration uses ElevenLabs on your key: 24+ voices with speed and stability control, and every scene duration-synced to the audio automatically.

Or record your own take and apply voice conversion, then clean it up with word-level trim, filler-word removal, and silence removal.

Transcription

Transcription runs on local Whisper — on our server, not a metered API. Word-level timestamps, filler-word detection, and code-switched Hinglish with romanized captions.

Cost to you: $0 per minute. There is no key to bring and no meter running.

Bring your own keys

Your key, provider list price, $0 markup

We never resell tokens. Keys are encrypted at rest, used only for your workspace, and billed by the provider directly to you.

$0markup on any model call — list price, always
0credit systems — no packs, no expiry, no rollover math
$0/mintranscription — local Whisper on our server
The full list

Model by model

What each model does, whose key it uses, and what it costs you.

ModelWhat it doesWhose keyCost to you
Gemini (text)Scripts, scene prompts, AI titlesYoursProvider list price — the free tier often covers it
OpenAI GPT (text)Scripts and scene prompts, as an alternative to GeminiYoursProvider list price
Claude (via MCP)Drives the studio from Claude Desktop, Claude Code, or CursorYour Claude subscriptionWhatever you already pay Anthropic — nothing to us
Gemini / Imagen (images)Optional AI scene stillsYoursProvider per-image price
OpenAI (images)Optional AI scene stillsYoursProvider per-image price
ElevenLabsNarration across 24+ voices, speed/stability control, voice conversionYoursProvider list price
Whisper (local)Transcription: word timestamps, filler detection, Hinglish code-switchingNone needed$0 per minute — runs on our server
Motion components & rendering100+ scene components, rendering, YouTube publishingNone neededIncluded
Questions

Keys, costs, and models

Why bring my own keys?

Three reasons. Cost: you pay each provider's list price and we add zero markup — we never resell tokens. Control: your generations run under your own accounts, and your keys are stored encrypted at rest. Honesty: with no margin on usage, our only incentive is to make the studio itself good.

Are my API keys safe?

Keys are encrypted at rest and used only for your workspace's generations. Workspaces are isolated tenants, so your keys, projects, and channel connections are visible to your workspace only.

Do I need all three provider keys?

No. Motion-graphic scenes, Whisper transcription, rendering, and publishing need no key at all. You need a Gemini or OpenAI key for AI scripts and optional scene images, and an ElevenLabs key for AI narration and voice conversion — or record your own narration instead.

What does a typical Short actually cost?

Script and scene prompts usually cost a few cents at most — Gemini's free tier often covers them. Narration for a 60-90 second Short is a small ElevenLabs charge on your key. Transcription, motion-graphic scenes, rendering, and publishing cost nothing. There is no ReelMint usage fee during the pilot.

Will ReelMint ever switch to a credit system?

No. Credits obscure what things cost and expire in ways that favor the vendor. Pricing stays your-keys-at-list-price for AI usage; when paid plans arrive after the pilot, they will charge for the studio, not for resold tokens.

See the whole pipeline on your own keys

Free founding-creator pilot. No credit card, no watermarks, no markup — ever.

Start free