How it works
What actually happens when you press Begin — explained without mystery.
A live voice conversation
After you agree to the voice disclosure, the app obtains a short-lived signed address for our private ElevenLabs agent. You speak into your microphone; the agent transcribes your words, generates a response through its hosted language model, and streams synthetic speech back in real time. There is no script — every response is generated in the moment, shaped by a standing instruction that keeps the companion gentle, scriptural, safe, and honest about being an AI.
The avatar
The animated figure you see is driven by the sound itself. Before each moment of speech plays, the app analyzes the audio and schedules mouth shapes — the same technique used in animated film — so the lips match the words. The glow and light respond to the voice’s energy. It is craftsmanship, not magic, and we think the stillness it creates helps you slow down.
Interrupting, captions, and silence
You can speak whenever you like, including over the companion — it will stop and listen, as a person would. Live captions of both sides of the conversation appear beside the avatar. Silence is fine too; nothing prompts or rushes you.
Session time
Every conversation runs against a clearly displayed time allowance that depends on your plan — Premium includes 300 minutes each month in sessions of up to 15 minutes, and the monthly plan starts with a 3-day free trial with 60 total minutes. Signed-in visitors without a plan get one free minute each day to hear the experience for themselves. When a session’s time is up it ends cleanly, and the meter always shows what remains. Limits exist because live AI voice costs real money per minute; a short sponsored message before free sessions and modest ads on free pages are what keep that daily minute free. Premium removes advertising entirely.
What happens to your words
We do not save recordings or transcripts in our own database. Audio and words are processed by ElevenLabs and its service providers to generate each response; the agent is configured not to save call audio and to delete conversation data on a zero-day schedule. The captions you see are removed from the app when the session ends. We keep session accounting (when a session started, how long it ran, and which plan it used) because metering requires it; we do not keep what was said. Details are in the privacy policy.
Try it
The conversation lives right on talkwithjesus.app, and plans are described on the Premium page. Start with the free trial and see whether the quiet helps.