Everything you'd ask before you sign up.

VidFlow is an autonomous AI pipeline that handles every stage of YouTube video production — from topic research and scripting through visual generation, voiceover, and final 1080p HD rendering. You bring the idea, the pipeline brings the rest.

One-prompt tools hand you a finished video and hope you like it. VidFlow runs a pipeline you can step into at any point — approve titles, lock the cast, swap a voice, re-roll a single sentence. Quick mode does the whole thing automatically. Director mode hands you the dial.

5 stages you see, 8 working steps under the hood, 10 internal states — the same pipeline viewed three ways. The 5 stages you direct in the workspace are Ideation, Script, Visual Bible, Voiceover, and Production — the 8 working steps (alignment and shot-planning split out, render and publish at the end) are how those tiles actually run.

None. The pipeline is automated. You guide creative direction through simple inputs and the system handles scripting, visual generation, voiceover, and composition.

Stickman, 2D Editorial, 2D Cartoon, Anime, Disney/Pixar, Realism, Dark History, AI Reconstruction. Each style has its own generation models and consistency rules.

Yes — every stage is editable. Rewrite scripts, swap visual styles, pick a different voice, split or merge shots, fine-tune the timeline before render. The seams are yours.

Re-roll just that shot. Every stage output is granular: one sentence, one shot, one voice take. You pay for the re-roll, not the whole render.

Yes. Channels are first-class. Each one has its own brand voice, visual style, voice defaults, and a private library. The workspace switcher scopes the entire app to whichever channel you are working on.

Yes. Open /try/director to drive a finished sample project stage by stage, or /try/quick-mode to see what a run costs. No login, no card — they run off a fixture, so nothing you poke at is saved. When you're ready, sign up and start your own.

Check spam first, then hit resend on the verify-email screen. The link expires after a while — if yours is stale, request a fresh one and the old link stops working. Still nothing after a few minutes? Email us and we'll verify you by hand.

Most of the time is the AI generation — scripting, images, then video clips per shot. The final stitch is a quick FFmpeg pass, usually 2–5 minutes once every shot is generated. Total wall-clock depends on length and how many shots there are.

Narration runs on neural voices — a searchable library with a play-preview on every voice. Want your own? Clone it: upload or record a short sample, or describe the voice in words, and keep up to 5 custom voices on your account. Cloning requires your explicit consent to use the sample.

You own everything you create. Project data is encrypted at rest and in transit. We never train on your content without explicit consent.

The first 350 credits are on us — no card needed. After that, four packs (Starter / Popular / Pro / Studio) priced per credit, billed once. No subscription required — though if you want a monthly cadence, the subscription tiers cover that.

No. Credits sit in your account until you spend them. There are no monthly resets and no use-it-or-lose-it deadlines. Top up when you need more.

Yes. Set up your channel profile with brand voice, visual style, and target audience and the pipeline tunes its output to match what already works for you.

The Enterprise subscription adds SSO and a shared visual bible, unlimited active projects, and a bigger monthly credit grant. No custom model fine-tunes — the pipeline runs the same models for everyone. ("Studio" is a one-time 15,000-credit pack, not a tier.) Talk to a human for custom pricing.