Talk to your own history.
Ask anything. Echo answers using only what you have actually lived, and shows you the experiences it drew from. It reflects what your record holds — it never tells you who you are.
Ask your history.
I answer only from what you've actually lived, and I cite the experiences I drew from. I reflect what your record shows — I never tell you who you are.
Deterministic and local — no model call, nothing leaves your device. Answers are grounded in your journal and pass the drift-guard, so Echo reflects, it never decides who you are. Ask a follow-up like “why?” or “tell me more” and it stays on the same thread.
How an answer is built
You ask; Echo listens to your journal
Every question is read against your own Experience journal — the things you lived, distilled into lessons. Nothing else is consulted. No model call, no outside knowledge.
It retrieves what you actually lived
Echo finds the experiences that speak to your question by meaning, weighted by what you've come to value. Topical questions need a real match; questions about the shape of your history fall back to the whole journal.
It answers only from what it found
The answer is composed only from the retrieved experiences, and every claim carries the ones behind it. A line with no backing is dropped before you see it. If nothing matched, Echo says so — it never invents a memory.
It reflects; it never decides
Every answer passes the drift-guard and ends by handing the verdict back to you. Echo shows you what your record holds — whether it represents you is always yours to decide.
Where it sits
vs. Ask the Mirror
The Mirror answers three fixed questions from your vault as it is now. Dialogue is open-ended conversation with your Experience journal over time — any question, always cited, always grounded in what you lived.
vs. /soul & /convergence
Those reveal the shape of who you're becoming — earned traits, a discovered through-line. Dialogue lets you interrogate the raw material directly: the individual experiences, in your own words, on demand.
vs. a chatbot
A chatbot answers from a model's training and will happily fill gaps with plausible fiction. Dialogue can only speak from your journal, cites its sources, and stays silent when it has nothing — honesty is structural, not a setting.
The sharp edges, shown
It answers only from what you lived
If your journal doesn't hold it, Echo won't say it. That means a real question can get an honest 'nothing here yet' — which is the point. It errs toward silence over a confident guess.
Retrieval is a keyword heuristic
Experiences are matched by concept, canonicalized and value-weighted, marked derived. It can miss an experience phrased in entirely different words. Better retrieval (embeddings) is a future upgrade; the grounding discipline is not.
The guard is patterns, not comprehension
The drift-guard refuses the four forbidden claims by structural rules, so Echo's answers are sovereign by construction. It can't catch every phrasing of overreach — but it makes 'never a verdict' a rule, not a promise.
Learned-from-others stays out
Only your first-hand experiences answer questions about you. What the Crew taught you (Law 12) is never folded into a first-person answer — second-hand wisdom is not your own history.