Settings

N8N Integration

Daily auto-content

Automation-ready architecture: generation is pure functions (bulkIdeas + buildRenderSpec) with duplicate detection (dedupeTexts), so any scheduler (cron, GitHub Actions, Inngest, Trigger.dev) can call it without code changes. Configure category, language, tone, count, templates and aspect ratio in the Bulk page.

AI providers & environment

• AI_IMAGE_PROVIDER — name of the image vendor (default: pollinations, keyless).
• NEXT_PUBLIC_AI_IMAGE_URL — override the image endpoint to plug OpenAI / Stability / Replicate behind your own route.
• Content understanding runs offline by default; swap RuleBasedContentAI for an LLM provider via the ContentAIProvider interface.
• No secret keys are shipped to the client. Uploads are validated (type + 10 MB limit).

Storage

Posts, custom templates, brand kit and analytics events persist in localStorage (entities: Post, Template, BrandKit, GenerationJob). Swap src/lib/store.ts for Prisma/Postgres when multi-user sync is needed — the UI only talks to the store interface.