Changelog

We've just released a new update that makes things a whole lot easier.

New dashboard

Cleaner home, faster brand switching from the sidebar.

Updated post creation

The flow is a lot clearer, and the new comment-to-edit system on posts is really intuitive — highlight a sentence and tell the AI what to change.

General interface updates

Tidier brand screens, smoother navigation, less hopping between pages.

Biggest update since launch. Three new surfaces, one rewired core.

A dashboard that actually shows you what's happening

The old dashboard was a list. The new one is a timeline.

• Vertical content timeline — every post, every brand, in one scroll. Granola-inspired.
• Content Plan — the whole month at a glance. Expand to multi-week view, collapse back to one viewport.
• Brand pips — every row tagged with the brand it belongs to.

Auto image model selection

Three image models live in Rheos now. Pick whichever you want, or let Rheos pick.

• Nano Banana 2 — fast, cheap, good for everyday graphics.
• Nano Banana Pro — higher quality, with thinking levels you can crank up for harder briefs.
• GPT Image 2 — slowest but smartest. Longer thinking, best brand fidelity, photorealistic results.

The model selector is inline on every generation. Each model has its own credit cost and ETA bucket so you know what you're trading off before you click.

Brand management hub

Identity documents are no longer static. They're a workspace.

• Tile-based layout — basic info up top, sources and socials side-by-side below, brand kit alongside.
• Click-to-document — every tile expands into the full document for that part of your brand.
• Scoped chat editor — a chat panel attached to your identity document. Edit your brand voice, sources, or audience by talking to it. Every change goes through an approval flow before it commits.
• URL-driven — Vercel-style brand-mode navigation. Switch brands without losing context.

Unified Post Editor — preview everywhere

• Per-platform publish previews — see exactly how your post will look on every channel before you ship.
• UPE modes — different layouts for different post types (idea, draft, scheduled, published).
• Slot-gen action — generate the next post in a series with one click.
• Best-times shared lib — schedule recommendations now use the same engine across the app.

Behind the scenes

• Per-model credit costs so pricing is transparent.
• Streaming progress on every image generation with per-model ETA buckets.
• Web search grounding on Gemini for fresher, more accurate copy.

This is the foundation we'll be building on for the rest of the year.

Two things shipped together: the latest image model and a smarter way to iterate.

GPT Image 2 — added the day it dropped

OpenAI released GPT Image 2 and we had it live in Rheos on day one.

• Photorealistic output — finally good enough for product shots and lifestyle imagery.
• Better prompt adherence — does what you ask for instead of approximating.
• Brand-true scenes — feed it your logo and it composes around it instead of erasing it.
• Pick it from the model selector on any image generation. Falls back gracefully if the API has a wobble.

Edit mode — iterate without losing the thread

Generated an image you almost like? Now you can edit it instead of starting from scratch.

• Lineage tracking — every edit is linked to its parent so you can compare versions.
• Reference-aware — keeps the parts you liked, changes the parts you didn't.
• Same model, same credits — just hit "edit" on any generation.

These two together mean you actually finish an image inside Rheos, instead of bouncing out to a designer to fix it.

After four months of building, Rheos is live and open to everyone.

What you get on day one

• AI content creation — text posts, images, and carousels, generated on-brand from your website.
• Scheduling across every major platform — LinkedIn, Instagram, Facebook, X, and Threads from one calendar.
• Brand learning — paste your URL, Rheos reads your site and learns your voice, audience, and visual style.
• Free tier — 5 posts a week, no card required.

This is v1 — the foundation. Everything from here is making it sharper, faster, and more useful.