This page records what is true right now. It should be updated whenever Lucia’s runtime, routing, or launch reality changes.
Live Transition Status
Lucia v0.1.3.6 remains the current Development baseline for Engine behavior, Admin UI, CI gates, and Eval Labs review. Eval Labs production routing is aligned to Development as of the correction live-verified on 2026-07-10; the earlier routing drift is recorded below rather than erased.
Current live-dev build identity:
The Admin/Engine build identity tag is runtime evidence for the currently deployed Admin/Engine pair, not decorative copy. Canon must not hardcode or fake deploy identity and should not record volatile commit hashes as doctrine.
Staging promotion is intentionally held until Eval Labs captures and reviews the v0.1.3.6 Development behavior baseline.
Current Version Stack
Current payment truth development proof snapshot:
Current Product Lines
HelloLucia, LLC currently owns three major product lines:
Current Canon boundary:
This product-line framing does not claim that guest-facing Lucia is fully production-hardened.
Active Public Surfaces
Development:
Current payment truth development surfaces tested:
Staging:
Evaluation:
Guest-Facing Lucia:
Marketing / Waitlist:
Documentation Library:
Documentation Library
HelloLucia, LLC documentation is published as one unified Mintlify library at helloluciallc.com. Its product switcher has three documentation homes: Lucia, Lucia Fieldwork, and Eval Labs.
This repository and its Mintlify deployment replace the two Obsidian Publish sites as the canonical documentation home. The legacy sites’ continued availability is not redirect proof and does not give them current source-truth status. See ADR - Unify Docs into Mintlify Library.
Current External Routing
Staging can receive these signals but is not targeted unless provider webhooks are manually switched.
The live Eval Labs app resolves to:
A Netlify redeploy is required after changing that Vite environment variable.
Confirmed routing history:
This establishes a continuous routing-configuration drift window. It does not prove that every individual historical run in that interval executed against Staging, and historical run records must not be rewritten to imply evidence they did not capture.
Eval Labs now exposes an owner-only Lucia Runtime Verification dashboard. It reads provenance from the same /admin/operator-focus response that produced the answer and shows the resolved endpoint, classified environment, Engine commit SHA, configured model, provider-resolved model, model-invoked state, deterministic-path state, fallback state, and verification timestamp.
Operating Loop
Lucia is moving away from localhost-first development.
Workspace OS Milestone
Lucia has crossed the Workspace OS / context-aware operator companion milestone in Development/v0.1.x.
The product transition is:
Current doctrine:
This is Development/live-dev runtime truth. It is not a production-readiness or staging-promotion claim.
Guest-Facing Lucia Milestone
Guest-facing Lucia has crossed from future concept into a distinct live-development product surface.
Current guest-facing runtime truth:
Lucia Engine and Lucia Guest Agent both currently use gpt-5.6-sol, but neither inherits configuration or runtime proof from the other. The Engine exposes same-response runtime provenance through /admin/operator-focus; the Guest Agent does not expose or persist that Engine provenance contract. The Guest Agent’s provider-returned model was verified independently on 2026-07-12.
Guest Agent v0.1.0 definition:
Versioning boundary:
Current guest-facing UI truth:
Identity orientation options:
Current guest verification and signal truth:
This is implemented live-development state. Runtime files, persistence, rate limits, replay/spam protection, Admin review/link tooling, and launch privacy copy still need hardening before production readiness is claimed.
Guest-facing Lucia also requires a first-class Guest-Facing Lucia Eval Track in Eval Labs, separate from operator-facing Lucia evals. Manual testing cannot cover every possible guest phrasing, so Eval Labs must stress-test guest identity, verification, privacy, tone, and guest-to-operator routing at scale.
Payment Truth Foundation
Lucia has completed the Development payment truth foundation across Stripe sandbox movement truth, Lucia Core durable ledger storage, LIEA payment attention suppression, and Admin read-only rendering.
Current implemented Development truth:
Tested live-dev proof:
Current boundaries:
See Lucia Payment Truth Foundation for the detailed payment truth architecture, route contracts, proof set, and deferred gaps.
Maintenance Intake
Waitlist Outbound Email
Current waitlist outbound email status:
Current verified behavior:
Known gap:
Current media pipeline:
Twilio media URLauthenticated Engine fetchBunny StorageBunny CDNAdmin UI render
Known correct CDN base:
Lucia Engine Model Layer
Current Development truth:
The stable top-level /admin/operator-focus contract remains ok / data / meta. meta.runtime reports Engine identity, while meta.model_runtime reports configured_model, resolved_model, model_invoked, deterministic_path, fallback_used, and per-task evidence. Model invocation and no-attempt deterministic execution are complementary states; fallback is recorded independently. Successful task evidence is required before provider-backed output is claimed.
Live owner verification on 2026-07-10 proved:
Lucia Guest Agent also currently uses gpt-5.6-sol, but through its own VV_AI_MODEL production override and source fallback. Its independently verified provider-returned model does not extend or reuse the Engine’s meta.model_runtime contract.
Model changes must preserve:
Current live-dev product architecture:
Current product pillar:
Current doctrine:
This is Development/live-dev runtime truth. It is not a production-readiness claim.
Current Admin runtime truth:
Current Engine runtime truth:
Repositories
Branches:
Admin Development CI:
Admin dev deploys after checks pass.
Current Deferred Items
See Also
