Introducing Skrivia — an AI writing, SEO & newsletter studio for Ghost creators

Integrations & API
1

Hi everyone,

I’m a Ghost creator myself, and over the last few years I’ve noticed something: while Ghost is an amazing publishing platform, most of us still rely on external tools to actually write, organize, optimize, and repurpose our content.

So I started building Skrivia — a modern writing, SEO, and newsletter studio designed specifically for Ghost creators.

:star: What Skrivia aims to solve

Across the forum I’ve seen many of the same pain points mentioned again and again:

  • Ghost’s editor can feel limited for long-form writing

  • Internal linking is manual and time-consuming

  • SEO requires jumping into multiple tools

  • Newsletters often need AI summaries / digests generated elsewhere

  • Repurposing posts for social platforms is a separate workflow

  • Managing multiple Ghost sites means switching dashboards

Skrivia brings these pieces together in one clean interface.

:star: What Skrivia includes

These are the initial features planned — all built directly around the Ghost Admin API:

:writing_hand: Clean Writing Experience

A focused long-form editor with better typography, outline support, version history, etc.

:link: Internal Linking Assistant

Search all your posts and pages instantly and insert internal links without leaving the editor.

:magnifying_glass_tilted_left: AI SEO Assistant

Automatically identify missing metadata, broken structure, opportunities for internal links, readability issues, etc.

:incoming_envelope: AI Newsletter Digest

Choose posts → generate a newsletter intro, summary, or digest using AI → publish through Ghost.

:recycling_symbol: Repurposing Tools

Turn any post into short formats for social platforms.

:star: Early feedback is extremely valuable

Right now I’m refining the first version and prioritizing features based on real creator needs.

If you’d like to follow the project or join the early access list, the landing page is here:

:backhand_index_pointing_right: https://skrivia.com

There’s also a short 30-second survey after signup to help shape the feature roadmap.

:star: I’d love your thoughts

  • What do you wish Ghost had natively in the editor?

  • What tools do you currently use alongside Ghost?

  • Would Skrivia help simplify your workflow?

Any feedback, criticism, or “this would be useful if…” ideas are extremely welcome.

Thanks!

Kevin

1 Like
2

Congratulations, looks nice.

I tried to use Ulysses before to be able to write on a native app on the go, but since I was not able to use builtin Ghost Cards, I always had to finish the article on Ghost editor anyway. That led to me that Ulysses just adds one more step to me, and Ghost editor is already good enough for a nice writing experience.

How do you handle Ghost Cards? How do you save articles to Ghost in the end? Do you just save as HTML? Are you save articles to Ghost as drafts? If so, what happens if you edit it on Ghost Editor and return back to your tool later?

3 Likes
3

Thanks so much, really appreciate your insight.

You’re absolutely right about Ulysses: without proper Ghost Card support, the writing workflow becomes fragmented and you end up in the Ghost editor anyway.

That’s exactly the gap I’m trying to address with Skrivia.

Here’s how I’m approaching it in the first versions:

1. Draft syncing via the Admin API

Skrivia creates/updates drafts directly inside Ghost.

The idea is to keep everything in sync so writers don’t have to juggle files or exports.

2. Respecting Ghost Cards

Full card editing is not planned for v1, but preserving existing cards and avoiding any destructive updates is a must.

Long term, I’d love to support the most common cards natively (images, callouts, bookmarks).

3. Handling changes from Ghost itself

If a writer edits the draft inside Ghost, Skrivia won’t overwrite blindly.

Detecting changes and pulling updates back into Skrivia is part of the workflow I’m refining.

Your comment is super helpful — you clearly understand the real-world edge cases Ghost creators deal with.

If you’re open to it, I’d genuinely love to learn more about the workflows you’ve seen with your clients.

Skrivia is still early, so expert feedback like yours really shapes its direction.

1 Like
4

From a technical standpoint, I’d be interested to know whether you’re interacting with Ghost by passing HTML back and forth, or Lexical. If you’re using Lexical, then you could 100% produce full card support. If you’re using HTML, be aware that there are multiple shortfalls, where exporting HTML and the re-importing it will mess up cards. (Visibility settings on CTA and layout of header cards come to mind, but I suspect there are additional examples.) If you want full built-in card support, you might want to yank logic straight from the repo and work with Lexical. If it needs to be HTML, then you’re going to need to write custom parsing on your end, and contribute patches to the places where html→card is currently lossy in Ghost, if you want to hit full card support.

2 Likes
5

That’s super helpful, thank you @Cathy_Sarisky for taking the time to write this.

Right now, my first iterations are based on working with HTML + the Admin API, mainly to validate the writing & SEO workflow, not to claim “full Fidelity” with Ghost yet.

But you’re absolutely right: if I want Skrivia to offer real, first-class support for Cards, HTML roundtrips alone won’t be enough.

The way I see it:

  • Short term: focus on creating drafts/updates in Ghost via clean HTML, while making sure I don’t break existing cards. That means being conservative about what Skrivia touches and where.

  • Medium/long term: proper Lexical integration is definitely on the table. If Skrivia is going to be more than “just another external editor”, then working directly with Lexical (or contributing to closing gaps in html→card) makes way more sense than trying to hack around it forever.

Your pointers about visibility/layout issues and lossy html –> card behaviour are exactly the kind of “gotchas” I need to be aware of before going deeper.

If you don’t mind me asking: have you experimented yourself with Lexical or custom tooling around Ghost’s editor? Any “don’t do this, you’ll regret it later” warnings are pure gold at this stage.

Thanks again for the detailed breakdown — this is incredibly valuable while I’m still shaping Skrivia.

6

Completely understood that an MVP might not have full functionality. That’s how MVPs work.

It’s a hallmark of AI use to ask for a high-effort reply to a low-effort question.

If you ‘update’ the body of a post that contains cards via HTML, you will mess the some cards up. You’ll need to test all cards and all settings on cards to determine which have full functionality, and ideally you’d do that before your users discover the problem on their own content.

2 Likes
7

Got it — that makes sense, and it’s a useful “do / don’t” to keep in mind.

Just for context: I’m French, so I do sometimes use AI to make sure my English is well-formulated :sweat_smile:

Not trying to push high-effort replies on anyone.

For the MVP I’ll keep things simple and avoid updating posts that contain cards.

Fair enough.

8

Using AI for help with a language you don’t speak is a great use! I do recommend asking the AI to translate, not compose, and I definitely recommend reading over it carefully to make sure it hasn’t done anything weird.

I delete a LOT of new posts from new users with 0 minutes on the forum and 0 posts read that are clearly AI written and that ask questions so vague as to be unanswerable. Pre-LLM, no one asked questions like that.

3 Likes
9

Small update on my side.

After a few conversations with Ghost creators and a short early survey available on the landing page, the same pattern keeps coming up around the writing → publishing workflow.

Most people don’t really struggle with writing itself (that’s not the point here right),
but with everything that happens around it:
SEO metadata,
internal linking,
newsletter prep,
and sometimes duplicating or adapting the same content in multiple places (sync matters ?).

Each tool used by the users who answered solves a small part of the problem,
but the overall workflow still feels fragmented for those who publish regularly.

That’s the area I’m focusing on right now, a frictionless workflow between writing, optimizing and publishing.