If you are a solo founder, you have already lost the battle with social media. Either you ignore it (and your launches die in silence), or you spend two hours a day reformatting the same idea seven different ways for seven different platforms (and you stop building product). There is no third option that works at the team-of-one scale.
Caster is the third option. You write one paragraph describing what you shipped, learned, or want to say. Caster generates a ready-to-post draft for every platform you have connected — formatted, sized, hashtagged, with platform-specific quirks already handled. You review, hit Post Now, and Caster ships them.
Today, Caster enters private beta.
What it actually does
Connect your accounts (OAuth, the boring part). Open Caster (it lives inside Konde Studio as a side panel; web app at caster.konde.io ships next month). Type the seed paragraph:
"Closed our seed round today — $2.4M led by Equation Capital. Going to use it to hire two engineers and ship the KSS App Store. Thanks to everyone who believed in us when this was just a Notion doc."
Caster generates seven drafts:
- X: punchy two-line opener, two sentences of detail, signoff. Within the 280-char limit. Hashtag-light because X has stopped rewarding them.
- LinkedIn: longer-form prose, opens with a hook, ends with a CTA. Around 600 characters. No emojis (LinkedIn's algorithm penalises them in 2026).
- Threads: Twitter-style but slightly more conversational. 500-char target.
- Bluesky: same target as Threads but less corporate, more "ship and tell" voice.
- Pinterest: short caption + hashtags + suggested cover image (pulled from your latest blog post automatically).
- Facebook: friendly, slightly longer than X, formatted for the LinkedIn-on-Facebook reader.
- Instagram: caption-only (Instagram does not let third-party tools post images), hashtag-rich, formatted for a carousel template.
You review. You edit. You hit Post Now. Done.
The voice
The thing that makes Caster useful is not that it can produce seven drafts — any LLM with a prompt can do that. It is that the drafts sound like you, not like marketing.
Caster reads your existing posts (we ingest your last fifty posts per platform when you connect) and infers the voice you actually use. Casual or formal? Long or short? Emoji-heavy or not? Indonesian-flavoured English or fluent corporate? Caster picks up on these and writes in your voice, not a generic one.
When the inferred voice gets it wrong, you nudge it with a one-line preference ("keep it casual, no emojis, max two sentences") that Caster respects across future drafts. Preferences persist per-platform.
Where it lives
Caster runs as a side panel inside Konde Studio v1.0+. You can also run it standalone via the web app at caster.konde.io (lands in private beta in April). The drafts go into a review queue you can act on from either surface.
If you have an AKK, you can also draft via API: POST /api/caster/drafts/internal. This is what our own agents use to draft posts on behalf of the team — the byline is yours, the writing is the agent's, you review before anything ships.
What it does not do
- Auto-post without review. Every draft requires human approval. We have hard-coded this and we will not relax it.
- Generate images. Caster generates captions and copy. Cover images are referenced from your existing assets (or from KW Render if you have it).
- Run analytics. Caster posts. The reading and engagement metrics live on each native platform; we are not trying to be Buffer.
These are deliberate scope cuts. Caster does the part that is painful (cross-platform reformatting and voice matching) and stops short of the parts that should stay human (review) or that other tools do better (analytics, asset generation).
Pricing
Caster is included in your Konde Studio license. No per-post fees, no monthly subscription, no upgrade screen. The only thing you pay for is your own LLM API costs — connect your OpenAI, Anthropic, or local model in Studio, and Caster uses your key. Typical cost per draft cycle (seven platforms): less than half a cent.
How to get in
Private beta runs through April. We are letting in roughly fifty Konde Studio users a week. Apply at caster.konde.io — there is a form, takes thirty seconds, we read every submission.
Public launch lands in May. If you cannot wait, write something interesting in the application's "what are you working on" field and we will probably let you in early.