Self-hosted, Self-Improving, Federated Everything App — dev management, agent orchestration, persistent mind, digital twin, second brain, and creative pipeline from a single dashboard over Tailscale to port the user into the OS
JavaScript ai-agentsdashboarddeveloper-toolsdigital-twinquantified-selfsecond-brainself-hostedtailscale
/**
* Where a project-specific jev head, its corpora and its cached embeddings live.
*
* One definition of each directory, in a leaf, because three separate modules
* need to agree about them and they disagree silently:
*
* `services/jevHeads.js` writes and reads the heads
* `services/jev.js` passes the heads directory to the sidecar
* `services/backup.js` excludes two of the three by path
*
* Kept OUT of `services/jevHeads.js` so the sidecar lifecycle — the module
* `jevRouter` defers specifically to keep the pinned model contract out of
* static import closures — does not have to import the head store (and through
* it zod) just to spell one directory.
*
* Everything under `jevDataDir()` is a derived record of this install's private
* repository history: machine-local, never federated, never in a status or
* capability payload. See the ADR
* [privacy records machine-local](../../docs/decisions/2026-08-08-privacy-records-machine-local.md)
* and the guard `services/sharing/jevNeverFederates.test.js`.
*
* Pure apart from reading the install root: no I/O, no process state.
*/
import { join } from 'path';
import { PATHS } from './paths.js';
/** The one directory every jev artifact lives under. */
export const jevDataDir = () => join(PATHS.data, 'jev');
/** Trained heads — adopted and candidate. Retained in backups: not regenerable. */
export const jevHeadsDir = () => join(jevDataDir(), 'heads');
/** Built training corpora. Rebuildable from the forge, so excluded from backups. */
export const jevCorporaDir = () => join(jevDataDir(), 'corpora');
/** Cached frozen-encoder outputs, per decision. Pure cache, excluded from backups. */
export const jevEmbeddingsDir = () => join(jevDataDir(), 'embeddings');
/**
* The embedding cache for ONE decision.
*
* Per-decision rather than one shared pool, so a training run can prune the
* keys it did not use without deleting another decision's cache. A flat pool
* would make that prune unsafe and the cache therefore unbounded.
*/
export const jevDecisionEmbeddingsDir = (decisionId) => join(jevEmbeddingsDir(), decisionId);