speak-it: quick spoken-word presentations
I wanted a simple text-to-presentation workflow: type what I need (a one-minute brief, a quick explainer, something to rehearse out loud), get a short script, then hear it instead of squinting at paragraphs. Speak It is that. The backend drafts the text, ElevenLabs reads it back, and I get a player in the browser. It is available at https://speak-it.samu.space/.
You pick English, Hungarian, or French, a tone (professional, casual, or storytelling), and a voice per language. Optional web search pulls in fresher facts when the topic needs it; optional image search adds a visual next to the result. The UI shows a rough time estimate before you commit, and generation streams in stages (search, script, voice).
When search is on, you also get sources you can trace. Outputs stay within a listenable length (think a few minutes, not a podcast episode). Repeat prompts can hit a fuzzy cache, and recent runs get shareable links via ?id= so you can send someone the same audio and text.
Stack-wise it is FastAPI behind the scenes and a React + Vite front end, plus OpenAI for the script (with optional web and image search) and ElevenLabs for TTS.
