> For the complete documentation index, see [llms.txt](https://lashae.gitbook.io/lashae-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://lashae.gitbook.io/lashae-docs/the-protocol/echoes.md).

# echoes

Two walkers, two sealed vows, one moment of resonance — anchored on Solana. No names, no plaintext, no social graph. Just proof that a pairing existed.

## The problem

Every social product treats "you're not alone" as a marketing claim. LASHAE does not want a claim. It wants a **receipt**.

If two walkers, on opposite sides of the world, sealed and revealed vows about the same quiet struggle — leaving a job, sitting with grief, returning to a practice — the field should be able to prove that pairing without ever introducing them or exposing what either one wrote.

## The protocol

1. When a vow is **revealed**, its plaintext is embedded server-side using `openai/text-embedding-3-small` (1536 dimensions).
2. The embedding is stored on `sealed_vows.embedding` (opt-in per walker).
3. A `pgvector` cosine-similarity search runs against every other revealed, opted-in vow — from any other walker — via the service-role RPC `find_echo_candidate`.
4. If the top match crosses the threshold, an `echoes` row is created with:
   * both `vow_a_id` / `vow_b_id`
   * both commitment hashes (sorted so `(a,b)` and `(b,a)` collide)
   * the cosine similarity and its **band**:
     * `near` — 0.82–0.87 — walking toward the same country
     * `close` — 0.87–0.92 — walking toward the same horizon
     * `mirrored` — 0.92+ — walking toward the same stone
   * `echo_hash = sha256("LASHAE_ECHO" ⑉ a_hash ⑉ b_hash ⑉ band ⑉ created_at_iso)`
5. Either walker may anchor the `echo_hash` to Solana via the Memo Program.
6. Both walkers see the echo in their field. Neither sees the other.

## What the public sees

The public receipt at `/e/:echoId` shows only:

* the two commitment hashes (of already-public vow receipts)
* the similarity band (never the raw score)
* the timestamp
* the echo hash
* the Solana signature (if anchored)

It does **not** expose walker ids, handles, embeddings, or plaintext. Exposure is enforced by the security-definer RPC `get_echo_public`.

## Privacy posture

* **Opt-in.** `sealed_vows.echo_opt_in` defaults to true but can be flipped from settings. Opting out clears the embedding — the vow is no longer semantically searchable.
* **Post-reveal only.** Sealed-but-unrevealed vows are never embedded.
* **One-shot pairing.** Once a vow is matched (`echo_id` set), it never echoes again. No promiscuous re-matching.
* **Service-role search.** The cross-user vector query only runs inside a server function under the admin client. Regular clients cannot query embeddings.

## Why this matters

The point is not matchmaking. The point is **proof of resonance without disclosure**. You can be met without being seen. The field does not name the walkers. It only proves the pairing existed.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://lashae.gitbook.io/lashae-docs/the-protocol/echoes.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.