Launch Validation

Investor Room / launch-validation

Launch validation · dated evidence snapshot, 26 August 2026 · the ledger and Atlas figures that are computed live are on the investor page and the Capability Atlas

Operationally live. Formally not yet a completed launch.

Talomnia is running in production, its investor documents and public execution evidence are live, and its current presentation passes the measured bilingual browser gates. The formal launch verdict remains NO-GO because the launch contract does not permit a historical P0 ordering violation to be rewritten as a pass, and two P1 deliverables remain open.

Operational status

LIVE

Formal launch verdict

NO-GO

Current criteria

34 pass · 3 fail

Commercial cases

0 · pre-commercial self-use only

What changed since the first report

This report supersedes the dated 19 August 2026 snapshot for current decision-making. The original report remains unchanged in the public trace. Its findings about missing investor documents, empty research, mobile overflow, weak accessibility coverage, incomplete Atlas presentation and undocumented knowledge export have since been remeasured and closed. Its historical findings about launch order remain part of the record.

P0 stop factors

P0 release gates

Gate	Verdict	Fail-capable evidence
Process gates (§14.1)	FAIL	The public trace preserves the initial design sequence; the later Editorial Evidence Atlas research precedes the redesign but does not rewrite the first sequence.
Honesty of presentation (§14.4)	PASS	Content gates reject invented traction, unlabelled targets, final commercial terms and unregistered-entity claims.
Publication Evidence Gate (§12.1)	FAIL — historical order	The complete self-use case exists now, but timestamped deployment and gate records prove that the first publication happened earlier.
Workflow artifact matrix, timing and cost	PASS	Every published workflow resolves to its evidence page; mutation tests remove source rows and observe the rendered change.
Workflow Ledger	PASS	Schema and cross-field validation cover hierarchy, time, cost, dispositions and rework links across 604 entries.
Public-repository sanitization	PASS	CI runs denylist and secret-scan gates with positive controls before public projection.
Request-data minimization	PASS	A mutation that routes submitted personal data into a persisted non-PII field is rejected by the database sweep.

Complete acceptance checklist

Each row names an instrument that can fail. “Pass” means the current measured release meets the criterion; it is not a claim of third-party certification.

§14.1 · Process gates

Criterion	Verdict	Check
Stage 0a and 0b validated before implementation	PASS	Repository and capability-discovery artifacts precede the implementation chain in the public trace.
Design research before design creation	FAIL — historical	The first design claimed three artifacts that did not exist. The later redesign corrected the method, not the history.
Sanitization policy checked in CI	PASS	Positive-control secret and denylist gates run in CI.
Stack, database and analytics ADRs precede main implementation	PASS	Committed ADR dates and dependency order are checked against the implementation history.
Only public, sanitized artifacts reach the site	PASS	Projection tests prove draft/internal records disappear from rendered output.
Git-to-database projection is one-way and reproducible	PASS	Bundle assembly pins Git objects; projection/preflight compare payload and database counts before every switch.

§14.2 · Site

Criterion	Verdict	Check
All required sections are reachable	PASS	The URL inventory and link-resolution sweep require HTTP success in both locales.
RU/EN content and terminology parity	PASS	The language gate reads both locale trees and fails on mixed-language or missing governed content.
Language preference works and persists	PASS	Browser tests exercise selection, storage and return navigation.
Theme preference, system theme and pre-paint behavior	PASS	Rendered tests cover system/light/dark state and the pre-paint script.
Mobile adaptation	PASS	Playwright geometry runs from 320px through 1920px and fails on page overflow.
Accessibility criteria	PASS	Keyboard, focus visibility, target size, reflow at 400%, text spacing, contrast and reduced-motion checks can fail independently.
Repeated entities render from the database	PASS	Source-row deletion mutations change cards, documents and workflow evidence.
Public caching; forms not cached	PASS	Header tests assert public cache policy and form no-store.

§14.3 · Evidence base

Criterion	Verdict	Check
Database-backed workflows with stable pages and resolving links	PASS	Workflow pages and outbound evidence links are swept as resolvable URLs.
Workflow 0 is a process, not a file	PASS	The launch workflow exposes stages, artifacts, failures, rework and result evidence.
Published market research and pagination	PASS	Research contains two projected, source-backed documents and the paging contract is mutation-tested.
Filterable Atlas, entity pages and graph relations	PASS	Capability Atlas exposes 115 entities and 223 projected relations with search and relation tests.
Investor documents, honest status, gates and CTA	PASS	RU/EN White Paper HTML/PDF and Pitch PDF/PPTX/HTML actions resolve from the compact document shelf.
Task and total time/cost	PASS	Ledger aggregation and rendered workflow totals share the same committed source.
Complete Workflow Ledger fields and rework links	PASS	Schema and cross-entry validation run over all committed ledger files.
Created/reused/modified disposition per artifact	PASS	Disposition-key parity is enforced and aggregate reuse accounting is generated from it.
Bilingual Pitch factual consistency	PASS	The four downloadable files share one 12-slide semantic contract; factual gates compare deck, White Paper and site.

§14.4 · Honesty

Criterion	Verdict	Check
No commercial cases stated plainly	PASS	The shared public status and investor narrative state zero commercial cases.
Targets are not traction or promises	PASS	Honesty gates reject unlabelled target figures.
No unsupported labor-cost comparison	PASS	Published-content sweeps reject the comparison.
Pre-seed is a hypothesis	PASS	The Investor Room and Pitch use the same hypothesis wording.
Research is marked pre-commercial self-use	PASS	The marker is required on projected research pages.

§14.5 · Integrations and infrastructure

Criterion	Verdict	Check
Forms reach Support Center and statuses synchronize	FAIL	Native forms and the sender are tested, but the production receiver path remains deliberately dark until receiver authentication and hostname constraints are completed.
Contract incompatibility blocks release	PASS	A vendored-contract mutation turns the suite red.
Idempotency and audit log are tested	PASS	Duplicate-submission and audit-transition tests exercise the database path.
Minimum request data stored	PASS	A personal-data persistence mutation is detected.
Production infrastructure and verified operations	PASS	Every release runs migration, projection, preflight, health and edge verification; rollback remains available.
SEO inventory	PASS	Sitemap, robots, canonical, hreflang and social metadata are swept.
Client knowledge-export procedure	PASS	The committed export procedure and its CI check fail when required inputs or boundaries are removed.

§14.6 · Ecosystem

Criterion	Verdict	Check
Talomnia is represented on arcanada.ai	PASS	The live RU/EN ecosystem inventory contains Talomnia and a resolving product link.
Arcanada blog article published through Publisher	FAIL	No publication receipt or live article URL exists. This outward publication remains operator-gated and is not fabricated here.

Measured execution boundary

On 26 August 2026 the committed ledger contained 604 entries: about 99.2 hours wall time, 78.8 hours active execution, 3.25 hours recorded human review, and 33 rework entries. The reported total cost is approximately $3,014.04, but it is an estimated, incomplete aggregate: cost is populated on 562 of 604 entries and human time on 567 of 604. These are self-measurements of the launch workflow, not customer economics and not evidence of savings against people.

Known limitations and next evidence gates

1. Formal launch history: the two P0 ordering failures remain in the record unless the operator records an explicit override; later product quality does not erase them.
2. Commercial evidence: there are still zero external paid cases. EG-1 is the first selected paid engagement; EG-2 is repeatability with acceptance and gross margin that includes human review.
3. Support delivery: the production receiver path must be enabled and proven end to end before the Support Center criterion can pass.
4. Ecosystem publication: the Arcanada article remains unpublished because Publisher dispatch is an explicit outward-action gate.
5. Payments and legal status: production payment acceptance and legal-entity registration remain external operator decisions. The site collects intent only; the legal entity remains pending registration.

Decision

Operational status: LIVE. Talomnia can be explored, its evidence can be reviewed, and selected design-partner / pre-pilot requests can be submitted without an online charge.

Formal launch status: NO-GO. The launch contract's P0 historical ordering failures have no recorded override, the Support Center production path is not end-to-end, the ecosystem article is not published, and no external paid case exists. This verdict does not take the live product offline; it defines what still has to become true before the epic can honestly be called a completed launch.