Eve library

Tara web search · delivery receipt · 8 September 2026

Guide

Revision 1

On this page

The sixteen-slide guide and PDF follow a query through the actual local web component. Two readable, continuous source captures, native inventory/history relationships, two branch diagrams and worked comparisons explain matching, recorded history and the supported navigation boundary. Authoring records source, sequence and optical decisions. Baseline: f1f956048cbf7a6a61c761eb9dc776e469bbfdfa; primary pin: 1ceae52c7896a59ac74a4429d410f3218f0b91c6.

The library has 67 guides, 722 unique slides, 745 guide-slide instances and 405 cited paths. All 706 earlier slides, scripts, narration entries and MP3 bytes remain exact. Sixteen new tracks add 1,200.563 seconds (20.0 minutes). No whole Tara source unit or full V1 workstream is closed by these citations.

Retained evidence#

  • Nineteen source cases execute the complete search component, route wrapper and unmodified helpers. The declared Next router records calls; it does not render destination pages. The component's 28 local examples and ten trending queries remain authored data. Pinned web/library tokens and Lilith CSS provide styling. Continuity binds all nine cited or loaded sources to the pin and baseline.
  • The browser review passes 117 checks with 64 captures, 48 embedded accessibility runs, five enlarged-text cases and three rejected removals: collections, result-history event and meditation no-push outcome. No fit findings or unexpected browser errors remain.
  • The focused center passes fourteen checks, preserves all 66 previous guide HTML files and four shared UI assets, retains 32 embedded slide captures, checks two audio boundaries and exports all sixteen PDF pages. Browser audio output was disabled. PDF text/order and the actual optical review provide distinct evidence.
  • The settled presentation suite passes 118 cases. New tracks receive encoded audio/pause checks and scoped decode. Speech reuse binds the exact prior scripts, entries and recording bytes. Freshness, coverage accounting, scoped ESLint, Ruff, formatting and local-link checks pass.

Corrections and limits#

All sixteen desktop draft pages were inspected. The original four-row capture was too small; the final slide uses one unbroken row and explicit matching terms. The inventory and recent-history slides now show native relationships. All sixteen 320-pixel drafts were checked, then the three changed drafts were checked again at desktop and narrow sizes. Final PDF and embedded images are recorded in optical-review.json, without claiming human acceptance.

Two initial source-fixture assertions ran before React committed a clear/remount; the harness now waits for state and mounts synchronously. No product behavior was changed. An initial narration invocation preceded rebuilding the generated slide inputs; the build now precedes the four serial speech batches. Failed attempt logs are retained without counting them as passing delivery.

Run python3 docs/presentations/presentation-center/verification/v1-tara-search-2026-09-08/verify-receipt.py from the repository root. Later library changes invalidate the frozen receipt; use the immutable delivery revision for historical checks. The full V1 program remains active. Human comprehension/listening, full application accessibility, live catalog/member behavior, destination rendering and clinical acceptance were not performed.

The normal commit hook formatted the new manuscript, draft and source-report JSON. Formatting continuity reconstructs the exact original bytes or compares their retained data; no source meaning or generated HTML, PDF, image or audio bytes changed. The final 117-check browser matrix was rerun to bind the formatted manuscript input. Its pre-hook report and captures remain in attempts/pre-hook-browser. Current center, PDF and narration evidence retains its exact artifact hashes.