The words and phrases you keep forgetting, in a story you'd read anyway.

Save one from any app on your Android phone. It comes back in tonight's story, written at your level, so the review happens while you read.

The Traces list in TraceSpark, filtered to Target First, showing Soslayar, Valija and Despacho with their English glosses and more below.
TraceSpark in Dual Sync reading mode, with the same word, soslayar, marked inside a generated Spanish chapter.
Words this reader saved, and the chapter written to bring one back.

How your captures become a story

Every Trace you save sharpens what the next chapter teaches you. TraceSpark is the input engine you run alongside your speaking practice, not a replacement for it. Built for strong B1 to C1 readers.

Capture it where you find it

Share from any app, or a photo of a sign or menu. The Trace you just saved, a word, phrase, grammar pattern or cultural note, is processed in the background. You never leave what you were reading.

Built from your own Traces

Your Traces come back as Sparks: stories and podcasts built around your interests and your gaps. Material tied to your own life is recalled better a week later than generic material², so relevance here is not a nice-to-have. Tonight's episode is built from the words your week produced, in whatever app you met them.

Real dialects, written in

Most apps read to you in one flat accent. TraceSpark writes the script in a real regional dialect, slang and idiom included, then voices it to match.

Pick a genre. It keeps going.

Pick a topic you like and the engine keeps producing more of it, episode after episode, at your level. Read it, listen to it, or keep the script on screen while it plays. Same depth either way.

Timed by FSRS-6

Reviews run on FSRS-6, the open-source scheduler Anki itself runs, targeting 90% recall. Published, peer-reviewed research, not a black box.

Who It's For

Three learners, three ways the engine adapts. Pick the one that sounds like you.

Why It Works

Read the words you actually meet.
Not a catalog everyone else is reading.

Mining native content by hand is real work, not laziness. TraceSpark does the mining for you. Every word or phrase you capture gets a due date of its own. The scheduler¹ tracks what is about to fade and decides which chapter it resurfaces in, so you are always reading the words you are actually forgetting.

The Old Way (Generic AI Story Apps)

"A word you saved inside their app can come back. The one you met in a message, an email, or an article on your phone cannot."

Every chapter is built from what you captured, anywhere on your phone.

TraceSpark weaves the words, phrases, grammar patterns and cultural notes you save from any app on your phone into a continuing series of stories and podcasts, at your level, so what you already met keeps coming back until it sticks.

Capture a word at work. It's in tonight's episode on the way home.

US$12.50/mo, later onFree until then. Price locked 12 months from your first payment.

Free to install. No pressure to commit, and no penalty if you stop.

Get it on Google Play

References

[1]

Ye, J., et al. (2022-2026). Free Spaced Repetition Scheduler (FSRS-6) open-source memory model and development benchmarks.

Verify Source
[2]

Self-reference effect in second-language vocabulary. Psychonomic Bulletin & Review (2025). Across two experiments, words encoded in reference to the self were recalled better immediately and one week later.

Verify Source