What is LuniSpark? (What)
LuniSpark turns a video or podcast transcript in SRT, WebVTT, or plain-text form into highlight quotes and ready-to-edit draft posts for Facebook, LinkedIn, and X. Quotes are rendered word for word from the cues you select, each with its transcript timecode attached.
When should you use LuniSpark? (When)
Use it after publishing a video, podcast episode, webinar, or lecture, when you need social posts that quote what was actually said instead of paraphrasing it from memory or inventing a plausible line.
Where can you use LuniSpark? (Where)
Open /lunispark/ in a modern desktop or mobile browser. The transcript is processed on your own device; nothing is uploaded, and no account is needed to use the beta. Open the tool or explore other Lunibyte guides.
How do you use LuniSpark? (How)
Import the transcript file or paste the text, confirm the content is yours to use, tick the highlight cues (or start from the labelled rule-based suggestions), review the verbatim quotes and their timecodes, then copy the drafts or download the Markdown export with citations. Evaluate the benefits in your own context; a local-first design does not replace your responsibility to review outputs.
Why use LuniSpark? (Why)
It keeps promotion honest and cheap: quotes can never be fabricated because they only render from your transcript, and the AI budget is zero by design because the planned intelligence, LuniSpirit, runs locally on your device instead of renting a hosted model.
Important limitations
The beta is draft-only and never posts for you. Highlight suggestions are rule-based, not AI, until LuniSpirit is ready; files are capped at 1 MB, a malformed timestamp rejects the whole file, and everything lives in memory only, so reloading the page clears the session.
Meet LuniSpark: Turn a Transcript into Social Posts, Right in Your Browser
Every video and podcast episode hides a second product inside it: the quotes. A sharp sentence at minute twelve, the moment the guest finally explains the thing clearly, the line your audience would screenshot if they ever saw it. The problem has never been recording those moments — it is the hour of scrubbing, copying, reformatting, and rewriting them into something postable. LuniSpark exists to collapse that hour into a few minutes, and to do it without asking you to upload your transcript anywhere or pay a monthly AI bill for the privilege.
LuniSpark is the newest Lunibyte tool, and it is free in beta. You bring a transcript — an SRT or WebVTT caption file, or plain text — and LuniSpark helps you pick the highlight moments and drafts Facebook, LinkedIn, and X posts built from the exact words in that transcript. Not paraphrases. Not plausible-sounding inventions. The words that were actually said, with their timestamps attached.
Why quotes must stay verbatim
Most AI content tools summarise first and quote later, which quietly reverses the trust order. A summary can be excellent and a quote can still be fabricated, because the model was never required to point at its source. LuniSpark takes the opposite position: a quote is only ever rendered from a cue in your transcript that you selected. If a sentence is not in the transcript, it cannot appear inside quotation marks in your draft. Every quote carries its timecode, and clicking that timecode in the transcript list takes you straight back to the cue it came from. You can verify every line in seconds, and so can anyone you share the draft with.
That design choice is deliberate, and it is the reason LuniSpark can be honest about what its beta does and does not do.
What the beta actually does today
Open LuniSpark and the flow is three steps. First, import: drag in a caption file or paste the transcript text, and confirm the content is yours to use — that checkbox is a hard gate, not decoration. The parser accepts SRT, WebVTT, and plain text up to 1 MB, and it rejects a file outright if a timestamp is malformed, because a mis-timed quote is worse than no quote. Second, pick highlights: tick the cues worth quoting, or press the suggestion button for a starting set. Third, take the drafts: Facebook, LinkedIn, and X versions built from your selected quotes, with character counters, copy buttons, and a Markdown download that keeps the citations attached.
The suggestion button deserves an honest label, so it has one, right in the interface: the beta's suggestions are rule-based, scored from cue length and repeated key phrases. They are a decent starting set and nothing more. They are not AI, and the page does not pretend otherwise.
Where the AI is — and why it costs nothing
The highlighting step is exactly where a local AI belongs, and we are building one: LuniSpirit, our own on-device agent, will read the transcript in your browser and propose highlights the way a careful editor would. It plugs into this same page through the same quote-provenance rules, so nothing about the trust model changes when it arrives.
LuniSpirit is also why LuniSpark will never send you an AI bill. The industry pattern is to rent a big hosted model, mark up the tokens, and pass you a subscription. We chose the other direction on purpose: intelligence that runs on the device you already own, at a marginal cost of zero. No API keys, no per-word metering, no uploading your unreleased episode to someone else's server so their model can read it. Your transcript stays in your browser's memory — reloading the page clears it completely — because that is where it belongs.
Who LuniSpark is for
If you publish long-form audio or video — a podcast, a YouTube channel, webinars, lectures, a company town hall — you already own the raw material for a week of social posts. LuniSpark is for the creator who wants those posts to quote the work faithfully, and for the small team that cannot justify another subscription for a task that should be simple. It is draft-only by design: there is no "connect your account" button and nothing posts on your behalf. You review, you edit, you post. The words stay yours because they always were.
Try the built-in sample transcript on the page if you want to see the whole flow in under a minute — then bring your own episode and see how many postable moments were hiding in it.
From the Lunibyte team