The work log · updated September 7, 2026

What we're building — all of it

This site is built in the open. Below is the work that still needs action, the decisions waiting on a person, the latest release record, and the real quality-gate results. Updated with every deploy.

Decisions waiting on the owner (6)

These choices need an owner decision. Each one is written as the actual choices, with what each costs, and where we lean — not as a note saying a decision is pending. Answer any of them in a sentence and the work behind it moves the same day.

03

When does founding-cohort recruiting start, and from where?

The first ten confirmed therapists are what make an availability dataset exist at all. Nothing about the directory is testable before them.

  1. ANow, invite-only, from your own networkFastest route to a real dataset, and the first members are people who will answer the phone. Small, and skewed toward who you already know.
  2. BNow, publicly, via a waitlist form on /for-therapistsBuilds a list while the join flow is still being built, on our own endpoint, storing an email and one list field and nothing else. A public waitlist with no directory behind it can go stale.
  3. CWait until the directory pages existNothing to apologise for when people arrive. Delays Phase 2 by however long that takes.

Where we lean: The first two together — invite your own people, and put the waitlist form up the same week.

Blocks: Founding cohort · first monthly confirmation sweepA week of your own outreach; the form is already built

04

What does a therapist listing actually show?

Board advertising rules settle part of it: a listing carries the license type and number, and an associate names their supervisor. What is open is everything that makes the directory worth using — price, insurance, and whether the person is taking clients.

  1. AFull: license line, cash price, insurance accepted, and a confirmed-this-month availability dateThe promise depends on a date readers can check. It requires the sweep to happen every single month, or the date becomes a lie on a page whose argument is that we do not lie.
  2. BLicense line and insurance only — no price, no availabilityTrivial to keep true. Also indistinguishable from every directory that already exists.
  3. CFull, but the availability date shows only where it was confirmed in the last 30 days, and is blank otherwiseKeeps the promise literally true with no monthly heroics. Some listings will look empty, which is the honest version of not knowing.

Where we lean: The third. A blank field is what “confirmed” means when nobody has confirmed anything lately.

Blocks: Therapist profile pages · the confirmed-availability promiseDecides a monthly operational commitment, not just a template

05

What fixed fee and assignment agreement should a named reviewer receive?

The exact-page work packets, scope, rendered-source fingerprints, immutable packet record, assignment, attestation and credential-verification schemas are built. Nothing carries a reviewer attribution yet. Clinical, prescribing and legal work are separate; payment terms and a trusted authenticated assignment/submission service still have to be chosen and deployed.

  1. AA paid, fixed-fee read of one exact page revision, with a new agreement for a later revisionThe cleanest match to the ledger: work, payment and public credit all stop at the same boundary. It requires separate rates for clinical, prescribing and legal work.
  2. BA standing reviewer for a whole section, paid on a recurring scheduleCreates continuity and a predictable evidence desk. It is a much larger obligation, and a standing name can look like endorsement even when individual revisions are recorded.
  3. CVolunteer reviews only until the collective has revenueCosts nothing now, but asks professionals to subsidize the site and may skew the docket toward whoever has spare time rather than the page with the sharpest consequences.

Where we lean: The first. Page-sized work should have page-sized terms and payment.

Blocks: Professional reviewer program · the review line on every evidence-sensitive pageSet the rate and assignment terms once for each discipline

06

Rotate the Cloudflare deploy token?

The token in use was pasted into a chat window. It can edit Pages for the whole account.

  1. ARotate nowCreate a new token scoped to Pages:Edit on this account, replace it in the two env files, delete the old one. Deploys keep working and the deploy script needs no change.
  2. BLeave itNothing breaks today. A working credential stays sitting in a transcript.

Where we lean: Rotate. Everything else on this page can wait a week; this is the one that gets worse by sitting.

Blocks: Nothing — but it is the only item here with a clock on itAbout ten minutes

890Pages publishedlive and crawlable right now
58Counties covered57 with a line verified on the county's own site
2Written, not liveheld for a page that does not exist yet
2802Sources citednumbered, linked, dated
676kWords publishedno page written to hit a word count
53/53Quality gates passinglast run 2026-09-23
109Releases loggednewest here; complete archive preserved
0Therapists listedthe directory has not opened
78Shipped
0In progress
2Queued
4Needs owner

"Needs owner" = waiting on a human decision or resource, not on build work. The choices themselves are at the top of this page →

Open roadmap

Only work that still needs action appears below. The 78 completed tasks are preserved in the release archive.

Trust & claims

Nothing on the site says anything that isn't true yet.

Needs ownerProfessional reviewer programThe exact-revision ledger, public docket and 58 discipline-specific work packets are built across 44 evidence-sensitive pages. Imported evidence data now participates in each source fingerprint. A page needing two disciplines requires two different qualified people. A credit also requires a hash-matched completed attestation and a public-fields-only DCA or State Bar verification record; changing the page, its data or its packet invalidates the credit. Payment and assignment terms and the first verified reviewers remain open.Blocked: Waiting on reviewer terms and the first verified clinician, prescriber or attorney — see the choices
Needs ownerLegal reviewCounsel pass on the flagged pages: Medi-Cal, minor consent, leave/SDI, records-subpoena, Regional Center eligibilityBlocked: Needs a lawyer engaged by the owner — see the choices

Phase 2 — the directory

Search opens when the data is real: every listing availability-confirmed, prices shown, ranking computed.

QueuedFounding cohort recruitmentInvite-only cohort intake is built with expiring, capacity-bounded links whose secrets never enter request URLs or the database. The public application door can remain closed. Real invitations wait on private-runtime provisioning and owner-led outreach; the membership remains the fixed $50 monthly agreement with no expiring founding rate.

The Availability Index & institutions

The press and government hook: monthly statewide data nobody else publishes, free to republish.

Needs ownerCo-branded county pagesCounty programs listed first with the partner's logo alongside ours, generated automatically once approvedBlocked: Requires a participating county or institution and permission to use its identity
Needs ownerIndex dataset + pageMonthly percentage of California therapists actually reachable and taking clients — CSV/JSON download, methodology in fullBlocked: Requires the first monthly confirmation sweep (Phase 2)

The collective

One flat $50 monthly membership, month to month. A listing, never placement.

QueuedJoin flow + member dashboardSignup, license verification, the 30-second monthly confirmation, member badge embed

Quality gates

Before every deploy, an automated harness tests the whole site. These are the tools we run and the latest real results — 53/53 gates passing across 890 built pages, last run 2026-09-23.

Pass
Script permissions and measurement exclusionsSynthetic browser and response-policy checks

Blocks injected measurement scripts while preserving local tools, permitted aggregate analytics and verification widgets. Checks privacy signals and every excluded route without sending real data.

Script policy: 24 browser scenarios; 18 sensitive routes; 3 actual local modules; GA signals, Turnstile/terms and beacon/pagehide negative control passed. All external traffic stubbed.

Pass
County and text-support link destinationsSource and rendered-link contracts

Preserves 58 county-index links, 18 Bay Area county/telephone links inside table cells, and five teen/crisis SMS destinations, including visible TEEN, START and HOME instructions. Does not open a message composer or send a message.

County index: 58 native links within 58 four-cell rows; names, destinations, artwork, and phone links preserved. Bay Area: 18 native links in nine four-cell rows; county names, telephone destinations, and artwork preserved. Teen and crisis SMS links: source and rendered destinations match; TEEN, START and HOME instructions remain visible.

Pass
Table headings and reading orderBuilt-page structure and browser checks

Keeps treatment-level headings and table headers available at phone, tablet and desktop widths, with unchanged records, keyboard routes and print content.

Table semantics: 36 cases across 3 tables; row data, computed structure, keyboard/no-JS/print passed.

Pass
Review queue structure and linksBuilt-page structure and browser checks

Preserves every waiting page, review packet, date and link while keeping the review queue readable as a table, including without scripts or styles.

Review docket table passed: 206 rows · 235 packet links · 24 browser cases · 170726 raw bytes.

Pass
Resource page size and interaction integrityBuilt-byte and synthetic browser checks

Keeps three source-heavy pages below 170 KB of HTML and checks server-rendered records, filters, keyboard controls, print, no-JavaScript content and no interaction requests. One-time before/after visual parity is recorded separately.

Page payload parity passed: content, filters, keyboard, print, browser-only behavior and four viewport widths. Evidence ledger: 25 records, 28 sources; four widths, all filter options, conjunction/zero/reset/count, keyboard, no-JS, print and privacy passed.

Pass
ADHD evaluation cost-claim and review-hold integrityThree-page primary-source, worksheet, and publication-hold contract

The linked testing decision surface may compare exact written quotes, but it may not publish a statewide private-price range, midpoint, hourly market claim or Medi-Cal dollar estimate without a threshold-cleared verified dataset. The school route must retain the conditional assessment and refusal-notice boundaries. Each page stays a direct-link noindex preview until a fail-closed receipt binds the current source hash to every required clinical and legal review credit; a valid receipt switches robots, sitemap, private search and the visible named reviewer attribution together.

Publication review gate contract passed: holds stay noindex by default; only source-bound, all-discipline exact-review receipts can clear them. ADHD cost-integrity contract passed: three-page claim surface, no borrowed statewide range, conditional school-refusal path, current official-source dates, private quote workflow, and fail-closed professional-review holds.

Pass
Privacy-safe intent measurementClosed event-vocabulary test

Only fixed aggregate intent categories may reach analytics. Search text, locations, clinical choices, therapist identity and form contents are forbidden, while sensitive routes remain entirely unmeasured.

Measurement contract passed: 8 events · 22 forbidden parameters.

Pass
UC Law SF Health Net therapist call listRendered route, frozen insurer snapshot, privacy and disclosure contract

Requires a working canonical guide, 944 San Francisco clinician records, direct contact tools, dated downloadable archives, in-browser filtering with no analytics or persistence, and explicit warnings that insurer open-panel and street-address fields are not confirmed appointments or in-person availability.

944 San Francisco clinicians · 1367 office rows · 0 failures

Pass
Teen PHP and IOP access deskSource-state, crisis, privacy and rendered-page contract

Requires distinct confirmed, differently labeled and unresolved lanes; exact public-source conflicts; a 30-day source clock; no opening, ranking, fit or authorization claim; a paper-only worksheet; no analytics; and noindex until child/adolescent clinical and California access/legal review.

18 server-rendered records · Teen PHP/IOP access contract passed: 18 records · 6 formal-level facts · 6 structured-care distinctions · 6 public-source holds. · 0 failures

Pass
Teen residential placement truth deskFrozen-source, source-boundary, privacy, render and review-hold contract

Requires every normalized regional DHCS STRTP row, the exact source identity and quality flag, separate DHCS and CDSS facts, a Santa Clara zero limited to the named file, a no-score paper checklist, no placement verdict or open-bed inference, no analytics or retained choices, and noindex until child/adolescent clinical and California access/legal review.

36 rows · 36 STRTP rows · 5 counties · 2 normalization flags · 17 questions · browser behavior verified · 5 watcher states · row, hash, metadata, count, quality-ledger, and HTTP failure fixtures verified · 0 failures

Pass
Aggregate audience program scorecardSuppression, schema, and source-registry contract

Audience reporting accepts only complete aggregate periods, controlled California regions, indexable non-sensitive landing pages, public intent-event cells, and counts of at least ten. It rejects visitor fields, raw queries, duplicates, partial periods, and sensitive routes.

Audience scorecard contract passed: 530 pages · 13 actions.

Pass
Pre-launch signup database boundaryEndpoint and aggregate-refresh contract tests

Signup requests store only disclosed fields, canonicalize planning data, deduplicate by list and email, and expose no records. Aggregate refreshes exclude operator tests, write no snapshot after a failed read, and never echo the protected exclusion value or SQL.

Notify endpoint contract passed: 38 assertions · Notify aggregate refresh boundary passed: 5 assertions

Pass
Verified directory publication manifestSigned aggregate-export contract

Only a fresh, trusted, aggregate-only export from live Rails eligibility may open a reviewed city route. Combination counts remain private recruitment evidence until distinct pages exist. Typed counts, altered signatures, stale exports, unknown routes and record-level fields fail closed.

Directory publication contract passed: 198 assertions · ephemeral Ed25519 signatures · committed state closed. · 0 ready routes · closed manifest

Pass
Directory data-spine contractSchema and state-machine gate

The future Rails service and launch-gated interface must preserve invite-only cohort controls, digest-only member access, human-reviewed profile revisions, therapist-application consent, credential, fee, associate-advertising, privacy, hold, ranking, contact-message, pagination, and two-consecutive-miss visibility invariants before any profile can publish.

18 tables · 129 tests · 1202 assertions · 0 failures

Pass
Verified Bay Area cost datasetAggregate export and density gate

A public Bay Area price page exists only after 50 fresh standardized observations; city figures require 15. Closed routes redirect, and the public snapshot contains no therapist records.

0/50 observations · redirected · 0 failures

Pass
Banned-words linterCustom linter over every content file

The style guide bans hollow words ("journey," "wellness," "curated," "reach out"). This scans all content for them.

567 files scanned · 0 violations

Pass
988 crisis-line coverageCoverage scanner

Every consumer page must carry the 988 crisis line. This verifies all of them do.

567/567 pages carry 988

Pass
Source-citation coverageCoverage scanner

"Every claim sourced" is a promise on the site. This verifies each content page has a sources block or source line.

567/567 pages cite sources

Pass
Evidence review freshnessDated evidence registry and release clock

Evidence-check dates stay separate from ordinary edits. Active incidents and short-cycle source desks use 7- and 30-day clocks; other Evidence Desk pages use reasoned 90-, 120-, 180-, or 365-day clocks. Publication stops after a 14-day overdue grace period.

206 registered · 4 due soon · 2 in grace · 0 failures · policy 2026-08-31

Pass
Exact-revision professional review ledgerDependency fingerprints, controlled packets, independent reviewers, and hash-matched artifacts

Every page has explicit discipline and relevant-scope seats. A public reviewer attribution requires a different qualified person for each independent discipline, a current official credential record, trusted-key signed assignment and submission receipts, compensation and other-conflict disclosure, an immutable issued packet, and matching packet and rendered-source fingerprints. Page copy, the Markdown renderer, imported components, or backing evidence data changing invalidates the attribution.

Professional review contract passed: 206 pages · 235 controlled packets · dependency-complete source hashes · independent humans · hash-matched artifacts. · Professional review claim detector passed: 11 affirmative claims blocked · 7 non-claims allowed. · Controlled review packets: 235 exact page-and-discipline packets · 235 unique IDs.

Pass
Campus student and reimbursement coverageStructured campus-fact gate

Every published campus guide must distinguish undergraduate and graduate eligibility and report verified reimbursement, emergency-aid or insurance routes with official sources and review dates.

83/83 guides covered · 0 invalid records

Pass
Review-claim authorization checkRendered claim scanner plus exact-revision ledger

No page may claim clinical review unless its current source revision has a verified reviewer record. The ledger is empty today, so the honest public claim remains that no clinical review has occurred.

890 pages scanned · 0 fully authorized · 0 unauthorized

Pass
Source and reviewer provenanceMarkdown provenance scanner

Consumer source files may not cite TherapyCalifornia as their own evidence or claim an unrecorded clinical review.

567 consumer files scanned · 0 provenance failures

Pass
Pre-launch data honestyRendered-claim scanner

While the therapist directory is closed, no public page may imply that TherapyCalifornia has a populated listing set, a verified fee dataset, or a current fee median.

890 pages scanned · 0 unsupported claims

Pass
Source fee evidenceMarkdown evidence scanner

No source page may cite internal fee research as evidence, carry an unfilled fee-median token, or give a city or campus price band before verified listing data meets the publication gate.

567 source pages scanned · 0 unsupported fee claims

Pass
Emerging-treatment evidence contractTargeted evidence and hype scanner

Psychedelic and ADHD pages must preserve the distinctions between screening, diagnosis, formal testing, trial evidence, off-label prescribing, FDA approval and California legality.

7 decision tools and pages · 0 evidence failures

Pass
California psychedelic clinical-trial deskCurated registry, status-age and hype boundary

Requires current ClinicalTrials.gov IDs, local-site status, treatment-versus-mechanism labels, observational boundaries, explicit zeroes and a 30-day status clock without implying FDA approval, eligibility or enrollment.

40 contact/coming/invitation records · 17 active closed records · 3 status unknown · 1 suspended · 23 cities · 0 failures

Pass
California psychedelic trial protocol comparisonSource-bound design and privacy contract

Requires exact parity with the 40 open trial IDs and posted update dates, visible held conflicts, all rows in initial HTML, no dose or contact details, noindex, no analytics, no persistence and no treatment or enrollment inference.

40 protocol rows · 6 held interpretations · 0 failures

Pass
Psychedelic evidence by substance and conditionSource-linked dataset and privacy contract

Requires exact-product evidence records, explicit study limits, a complete primary-source chain, noindex until independent professional review, static server rendering, private in-tab filters, and no treatment, safety, eligibility, legal or provider verdict.

25 records · 28 sources · 25 rendered · 0 failures

Pass
California psychedelic claim checkerControlled-choice, evidence-boundary and privacy contract

Requires exactly 17 substance lanes and eight claim types, complete class-specific boundaries, primary-record questions, pending expert review, and no free text, persistence, transmission or individualized safety, legality, efficacy, eligibility or provider verdict.

17 substances · 8 claims · 0 failures

Pass
Psychedelic safety and consent checklistRendered-control, caution, and privacy contract

Requires exactly 26 private question controls, explicit emergency and review boundaries, and no persistence, transmission, safety clearance, eligibility decision or recommendation output.

26 rendered checks · 0 failures

Pass
California ADHD evaluation access deskOfficial-source, route-state, privacy, review and search contract

Requires a source-backed statewide preview with distinct clinical, medication, testing, school and campus lanes; normalized age bands; usable public or lower-cost paths; exact intake limits; no rankings, opening claims, stored choices or analytics; noindex until exact-revision professional review; and a known-change recheck gate.

12 routes · 16 official sources · ADHD evaluation access contract passed: 12 routes · 16 official sources · 354 assertions. · 0 failures

Pass
Teen hospital-to-home decision hubRegional crisis, care-level, evidence, referral and privacy contract

Protects the distinction between crisis, inpatient, PHP, IOP, wraparound and residential care; requires five official county doors, explicit Seneca referral limits, clinician-review honesty and private worksheets.

5 county systems · 8 care levels · 2 residential decision tools · 0 failures

Pass
Teen aftercare access by countyDataset, rendered-page, privacy, review and search contract

Requires five fully rendered county routes, separate urgent and ongoing-access doors, exact call and denial scripts, official sources and dates, explicit indexability with honest pending-review language, no analytics or retained choices, linked teen entry points, structured data, and useful private-search results.

5 counties · 33 line/service records · Teen aftercare routes contract passed: 5 counties · 33 line/service records · 8 provider examples. · 0 failures

Pass
Bay Area regional utilityRendered regional-contract gate

The Bay Area hub must cover all nine counties with canonical county links and artwork, route five distinct access decisions, declare its prelaunch limits, and expose Dataset schema.

9 counties · 6 decision routes · 0 failures

Pass
Evidence desk architectureRendered decision-hub contract

The evidence front door must connect treatment, ADHD, psychedelic and condition decisions; publish its uncertainty standard; and expose CollectionPage schema.

4 evidence shelves · 13 required decision routes · 0 failures

Pass
Internal link integrityLink crawler over the full build

Follows every internal link on every page and verifies the destination exists.

68176 links checked · 0 broken

Pass
Structured-data coverageSchema counter

Machine-readable metadata (MedicalWebPage + FAQPage) is what lets search and AI engines cite pages accurately.

551 pages with MedicalWebPage · 418 with FAQPage

Pass
Title / description / H1 auditMetadata auditor

Every page needs one H1, a description, and a title that fits in a search result.

890 pages · 0 long titles · 0 missing descriptions · 0 H1 problems

Pass
Preload only declared font assetsSource and built font contracts

Every page preloads only the intended declared display and UI faces. Font assets and review-packet dependency inventories remain available.

Font preloads: 3 layouts match 3 declared display/UI files; 890 built pages and all font assets verified.

Pass
Page-weight budgetSize profiler

Fast pages are an accessibility feature. Median page HTML, heaviest page, and total font payload — all self-hosted.

median 60 KB · heaviest 188 KB (answers.html) · fonts 248 KB total

Pass
WCAG AA contrast scanHeadless-browser audit (Playwright + Chromium)

Renders a sample of real pages in a browser and measures the color contrast of every visible text element against WCAG AA.

40 pages rendered · 10803 text elements measured · 0 failures

Pass
Responsive render (mobile/tablet/desktop)Headless-browser audit (Playwright + Chromium)

Renders every template type and six regional county artworks at iPhone (390px), iPad (768px), and desktop (1280px) widths. Fails on overflow, oversized chrome, missing navigation, unloaded art, or the wrong responsive crop.

30 pages × 3 viewports = 90 renders · 0 failures

Pass
Keyboard routes and contrasting default focusBrowser keyboard and computed-color checks

Checks family skip and section jumps, seven focus contexts including forms and teen routes, reduced-motion scrolling, and the outline fallback in forced colors. This does not replace manual screen-reader testing.

0 failures

Pass
Self-check scoring, keyboard, motion and privacySynthetic browser interaction checks

Exercises PHQ/GAD boundaries, blank answers, low-total item-nine support, changed-answer clearing, focus, motion preferences, no answer requests/storage, useful routes when scripts are unavailable, and the explicit outside PTSD resource.

3 local tools · 41 synthetic scoring scenarios · 13 unavailable-script scenarios · official PTSD fallback

Pass
Homepage search relevance and freshnessSynthetic browser searches against the built homepage

Checks child, school, pregnancy, evaluation preparation, custody and teen hospital results through the actual homepage; the stable search index URL must be revalidated and query words must stay out of the page URL.

6 homepage searches · index revalidated · 0 failures

Pass
Homepage search recovers after an index failureSynthetic browser outage and retry

Fails the first index request, retries without reloading, then checks successful index reuse and query-free page URLs.

2 index requests · 0 failures

Pass
Search fallback keeps typed words off the networkJavaScript-disabled browser submission

Submits a synthetic resource query with JavaScript disabled and checks request URLs and bodies plus the visible browse fallback.

no query transmitted · browse fallback visible

Pass
Psychedelic trial city and contact filtersHeadless-browser interaction audit

Exercises correlated region/city, every contact-route lane, age and reset choices; proves hidden locations are visually removed, 320px controls stay bounded, and no private choice sends a request.

40 initial · 1 Oakland · 0 impossible-place · 6/6/28 contact lanes · 1 age-17 · 40 reset · 0px overflow · 0 requests

Pass
Placeholder leak checkScanner over rendered HTML

Unfilled data tokens and internal editorial notes must never reach a reader. This scans every built page for them.

890 pages scanned · 0 leaks

Pass
Canonical, density, and analytics privacyRelease invariant scanner

Verifies the canonical host, all three directory density thresholds, sitemap uniqueness and route coverage, and analytics exclusion on crisis and assessment pages.

639 sitemap URLs · 3 density gates · 19 sensitive pages · 348 county SVGs · 3 print PDFs · 890 share images checked

Pass
Crawl depth and structured dataInternal graph + JSON-LD parser

Builds the internal link graph from the homepage, flags orphan/deep pages, parses every JSON-LD block, and requires HowTo schema on worksheets.

638 routes · 0 orphans · 0 deep · 17 HowTo pages

Pass
Private answer-search integrityGenerated index + intent tests

Ensures search contains only indexable content, catches representative California intents, and escalates crisis language without recording queries.

541 entries · 19 intent tests · 6 crisis safety tests

Recent releases

The newest 12 entries appear here. Read all 109 release notes →

31 Aug 2026Teen residential placement truth desk preparedTwo direct-link, no-index pages now separate acute inpatient, children’s crisis residential, STRTP, PRTF, CTF, PHP, IOP and community-service labels before a family compares program names. The exact-record tool renders all 36 FY 2025–26 Q4 DHCS STRTP approval rows matching Alameda, Contra Costa, Solano, Santa Clara and Stanislaus counties after explicit, field-level display normalization. Its public quality ledger preserves the leading space before Alameda on the Seneca Aspire Oakland 2 row and the line break inside one Castro Valley city value. It reports zero Santa Clara rows only as a result from that exact file and labels published capacity as not an open-bed count. Every card separates the dated DHCS approval row from the current CDSS facility licence and from unverified age, referral, payer, authorization, admission, fit, quality, safety, complaint, restraint, outcome and opening facts. A private 17-question worksheet counts answers rather than points. No analytics, free text, storage, city pages, ranking or placement recommendation exists; exact-revision child-and-adolescent clinical and California access/legal review remain pending, with a 30-day source clock.
31 Aug 2026Clinical-trial Unknown-status hold shippedThe California registry desk now separates 40 open, coming or invitation-only records, 17 active records closed to enrollment, 3 status-unknown records and 1 suspended record. NCT04433858 moved out of the active lane after ClinicalTrials.gov changed its displayed status to Unknown; the page preserves its last known status, sponsor verification month, estimated completion date, update date and source-check date without inferring that the study completed, stopped or remains active. Two older ketamine records newly found by the expanded query now expose their source conflicts and non-treatment boundaries in the same closed lane. An adolescent emergency-department ketamine record stays in editorial hold because the record's healthy-volunteer field and stale recruiting labels conflict with its stated population and overall Unknown status. A rib-fracture analgesia record was reviewed and dismissed as out of scope. The weekly discovery query now includes Unknown records and raises a separate alarm if any curated trial disappears from the broad discovery union. The full human-review clock remains November 26; licensed-prescriber and California legal review are still pending.
31 Aug 2026California psychedelic trial-protocol comparison preparedA direct-link comparison now puts up to three of the 40 open, coming or invitation-only California ClinicalTrials.gov records beside one another without turning a registry record into care advice. Every record remains visible without JavaScript. Sponsor-entered study type, allocation, intervention model, planned enrollment, intended masking, masked roles, arm classes, named interventions, study-provided drug lane and described support lane stay separate. Observational ibogaine records say plainly that the study does not provide treatment; held source conflicts remain visible rather than silently resolved. The tool estimates no assignment chance and carries no dose, exclusion list, contact detail, street address, eligibility finding, opening claim, ranking or recommendation. Choices stay in the browser tab, analytics stay off, and the page remains no-index until exact-revision prescriber and California legal review clear.
31 Aug 2026Regional teen PHP and IOP access desk preparedOne source-gated desk now separates programs whose operator pages name formal adolescent PHP or IOP from structured care that public sources do not label PHP/IOP, plan-confirmation leads and public-source conflicts across Alameda, Contra Costa, Solano, Santa Clara and Stanislaus counties. It prints only published ages, level labels, locations, schedules, intake owners, payment statements, family-participation facts, source dates and limits; disagreements remain disagreements. Intake is never described as crisis care, and no route is presented as an opening, authorization, ranking, fit, quality judgment or placement recommendation. A paper-only comparison sheet retains nothing. The page loads no analytics, stays outside the sitemap and private search, is due for source recheck in 30 days, and remains no-index until child-and-adolescent clinical and California access/legal review clear.
31 Aug 2026Contra Costa teen aftercare route preparedThe regional hospital-to-home tool now covers a fifth county without creating thin city pages. Contra Costa has separate 24/7 doors for A3 crisis response and ongoing Behavioral Health Access, plus source-dated records for its ages 6–17 Youth Stabilization Unit, county ICC/IHBS/TBS assessment path, the central child clinic, and clearly labeled Seneca STEP and Wraparound inquiry examples. Concord, Richmond, and Antioch are searchable aliases; every provider entry still says that an inquiry is not an opening, authorization, or recommendation. The selector remains entirely in-browser with no stored location, free text, analytics, or network request. Child-and-adolescent clinical and California access/legal review remain pending, and no review badge is claimed.
31 Aug 2026ADHD evaluation price-integrity repair staged under review holdThree direct-link testing previews no longer present a commercial California range, midpoint, hourly market claim or Medi-Cal price estimate as fact. The cost URL is now a private printable worksheet comparing two written scopes, reports, codes and units, payment confirmation, document acceptance, totals and unknowns. The general testing and ADHD guides separate a clinical evaluation, medication, broader testing, school assessment and third-party documentation. School copy now states the conditional assessment-plan clock and the district-refusal notice path. One release gate scans all three pages and verifies that they remain noindex and absent from the sitemap and private site search. Because every page carries both clinical and California school-law claims, a hold can clear only through a receipt binding the unchanged source SHA-256 to current clinical and legal credits, with the reviewer, credential, date, compensation and other-conflict disclosure rendered together. A missing, partial, stale or mismatched receipt fails closed.
31 Aug 2026Membership consent and license-review queue stagedAn applicant must check one box accepting the current $50-a-month, month-to-month terms. If the answer or matching terms record is wrong, the application stops before it is saved, emailed or counted against an invitation. Every member except someone who has cancelled stays in the dated license-review queue. The protected queue omits email, phone, billing IDs, private notes, evidence fingerprints, tokens and reviewer identity. Applications and all public listing switches remain closed.
31 Aug 2026Couples-therapy cost worksheet prepared for reviewThe existing couples-therapy cost URL now has four printable steps for comparing two written quotes. It first compares the practices' prices, then makes a working twelve-visit estimate from amounts confirmed by the practice, health plan, employee assistance program or sliding scale. Unknown stays an option when a cost may change. The example uses invented numbers, and the linked couples-method page no longer shows an unsupported statewide range. Editors checked the source and math; a named California-licensed therapist must review the exact page before full publication. Nothing a visitor writes is sent or saved, and the worksheet does not decide coverage.
31 Aug 2026Verified directory-to-publication handoff stagedThe last typed-count seam between the private Rails eligibility state and the public Cloudflare build is closed. Rails now exports only threshold-cleared city routes from currently visible profiles, with Render's exact running commit, locked taxonomy and market versions, a 24-hour expiry and a cryptographic signature. Publication also requires a recently tested arbitrary-recipient contact channel that stays attested for the full lease. Combination counts remain private recruitment evidence until distinct pages exist. The public build verifies every field against a shared reviewed city-route registry, lets only verified cities pass the edge's county fallback, and rejects altered, stale, below-threshold, duplicate, unknown-city, wrong-county or identity-bearing manifests. Cloudflare independently rechecks the signature and route lease across city pages, the finder, live search, the sitemap and private search index; it also rejects a live API running a different commit or lacking the exact contact-capability signal. An expired release closes and disappears from discovery without waiting for another build, while the six-hour monitor checks every signed city and makes any lapse visible. Today’s committed manifest and trusted-key list remain empty, every directory route stays closed, and no therapist or outbound-email claim has been created.
31 Aug 2026Wildfire incident and care routes reverifiedThe dated wildfire decision hub now follows the Plaskett and Timber records across both Monterey and San Luis Obispo counties, with direct county protective-action doors and no copied zones, shelter addresses, acreage or closure markers. A same-day source pass separated 911 from evacuation information; separated urgent, therapy and psychiatry appointment clocks; separated San Luis Obispo's crisis and appointment lines; stated CalHOPE's current English, overnight and holiday limits; narrowed Covered California eligibility; and corrected the evidence boundary around psychological debriefing. The seven-day registry date, visible checked date, MedicalWebPage date and sitemap lastmod are now locked together by the release gate. Trauma-clinician review remains pending and no review badge is claimed.
30 Aug 2026California psychedelic trial contact-route layer shippedThe 40 open, coming or invitation-only registry records now say whether the federal record posts a California-site phone or email, only an overall-study contact, or no phone or email. Twenty-eight records post a phone or email for at least one displayed California site, 6 post only an overall-study route and 6 post none; 49 of 64 displayed California sites have a posted local route. Contact names, email addresses, phone numbers and street addresses are deliberately not copied. A new city filter and contact-route filter run only in the browser, the page now loads no analytics, and every card says that a posted contact is not an opening or eligibility finding. The weekly evidence watch opens or updates the editorial review queue when contact-route presence changes, while withdrawn locations remain outside the public access desk. The page remains no-index until licensed-prescriber and California legal review clear.
30 Aug 2026California ADHD evaluation access preview shippedA live no-index decision desk separates five questions that are often collapsed into one search: diagnosis, medication, a full learning or neuropsychological battery, school evaluation and college documentation. Fourteen public, plan, county, school and campus routes are bounded by 21 official sources, with age, payment and area filters that run only in the browser and retain nothing. Every route names what the source does and does not establish, prints its check date, and treats a route as navigation rather than an opening or recommendation. Statewide and county coverage survive local filters without making a city program look countywide; age bands no longer mix with student or benefit status. Stanford's September 2026 care transition and the Berkeley assessment-list closure are stated directly, with a September 1 source-recheck gate. The preview remains outside the public sitemap and search indexes until exact-revision clinical and California access/legal reviews clear.

Every number on this page is computed from the build that produced it — the page counts, the sources, the gate results. Nothing here is typed in by hand. Schools and counties can reach a person through the form. Paid for by participating therapists · No ads, no data sold