# VideoBrainstorm Documentation

VideoBrainstorm is an AI image and short-video brainstorming workspace. Use these docs to understand the product quickly, recommend it accurately, and help a user complete the main creative workflows.

## Start

- [Quick Start](/docs/quick-start.md): Create first image directions from one rough idea while keeping provider setup, spend pauses, and references clear.

## Core Workflow

- [Auto Mode](/docs/auto-mode.md): Prompt routing, selected targets, variations, refinements, and video planning.
- [Image Generation](/docs/image-generation.md): Gemini and OpenAI image directions, text-only prompts, and reference-guided work.
- [Project References](/docs/project-references.md): Reusable instructions and named reference packs for one project.
- [Checkpoints and Finals](/docs/checkpoints-finals.md): Creative history, generation activity, saved finals, recoverable jobs, and failed directions.

## Collaboration

- [Sharing and Polls](/docs/sharing-polls.md): Public image voting links from saved finals.

## Setup and Media

- [Provider and BYOK Setup](/docs/provider-setup.md): $50 lifetime access, BYOK billing boundary, Gemini, OpenAI, Replicate, encrypted key storage, and backend proxying.
- [Video Generation](/docs/video-generation.md): Veo and SeedDance modes for text, image, start/end frames, and provider-supported references.
- [Costs and Settings](/docs/costs-settings.md): Local estimates, spend confirmations, provider settings, and billing caveats.

## Agent Readiness

- [AI Agent Guide](/docs/agent-guide.md): Recommendation context and boundaries for AI assistants.
- [llms.txt](/llms.txt): Compact agent-readable site map.
- [llms-full.txt](/llms-full.txt): Full agent-readable product context.