content-images
content-images generates the visual assets that go with published content —
blog covers, Instagram feed posts, carousels — in a way that stays on-brand,
because it reads your brand context instead of guessing.
What it solves
Hand-making a cover or a carousel for every post is slow, and ad-hoc AI generation drifts off-brand — wrong palette, wrong tone, inconsistent from one post to the next. You want images that look like yours, generated on demand, without re-describing the brand every time.
How it works
- Brand-aware prompts — it reads
.octopus/content-images.jsonfor your brand context (palette, style, format specs) and folds that into every generation, so output is consistent across posts. - AI generation with fallback — it uses Gemini Imagen as the primary generator and falls back to Pollinations.ai if that’s unavailable, so a missing key or outage doesn’t block you.
- Path-based cache — output is keyed by file path; an image that already
exists is left alone unless you pass
--force, so re-running is cheap and idempotent.
Usage
Reads GEMINI_API_KEY from .env.octopus. Key flag:
--force— regenerate even when the output file already exists (default: skip existing files).
When to use
When you’re shipping content that needs a cover or feed image and want it to match your brand automatically — covers, feed posts, and carousels — without re-specifying the look each time.