Throughline — therapist-facing service (plan)

Brand: the public product is Throughline (by HiveJournal) — "The throughline between sessions." It was codenamed JQ Connect internally; "Throughline" is the name therapists/clinics see, chosen so it scales from a solo practitioner up to a health-system integration without reading as a consumer add-on (clinicians don't know the "JQ" brand — their clients do). Brand architecture: one product, tiered (solo therapist → practice → health system buy the same thing at different seat counts) — not a sub-brand per segment. JQ Bridge stays the named under-the-hood sharing engine. OG share card: apps/frontend/src/app/api/og/throughline/route.tsx.

Thesis: therapists are a high-trust acquisition channel — a recommendation from your therapist converts far better than an ad, and it creates a retention loop (you keep journaling because your provider is in it with you). The unlock is metadata, not content — providers see trends, never entries. And we build it on JQ Bridge, which already does the hard part.

Naming: JQ Bridge is the underlying primitive (a user shares journaling metadata with a trusted contact). Throughline (codename JQ Connect) is the provider-facing product built on top of it (a therapist's lens: client roster, between-session vitals, engagement alerts). One sharing engine, a professional surface.

The big unlock — JQ Bridge already does most of this

Already shipped (jq_bridge_connections / jq_bridge_permissions, supabase/migrations/039_add_jq_bridge.sql):

• Metadata-only sharing — permission toggles for mood_trends, topics, patterns, insights, summary, recommendations. Raw journal content is never shared. This is the trust foundation and it's done.
• connection_type already includes 'healthcare' and 'professional' — the schema was built with therapists in mind.
• Consent + control — invite → accept, per-permission toggles, revoke any time (DELETE /api/jq-bridge/connections/:id). Sharing is one-directional, sharer-controlled.

Implication: a client can share journaling metadata with their therapist today. The Phase-1 funnel needs almost no engineering — it needs positioning, a landing page, and therapist outreach.

Compliance posture — decide this first (it shapes everything)

Recommendation: launch as a wellness / engagement tool, not a clinical one.

• Metadata-only, client-initiated consent, revocable, no diagnosis/treatment claims, no "risk detection."
• Alerts are engagement nudges, not clinical alarms. "Your client hasn't journaled in 14 days" / "mood trend dipping — might be worth a check-in." The provider decides what to do. Explicit disclaimer: not a monitoring or diagnostic device; does not detect crises. (Same wellness-not-clinical line we drew for the RCT — keeps us clear of FDA SaMD + liability.)
• The HIPAA / BAA fork: the moment a provider uses this in documented clinical care and we transmit identifiable client health-signal data to them, we likely become a Business Associate needing a BAA + HIPAA-grade handling. Don't trigger that on day one; design toward it. Add a provider_access_log (who-viewed-what-when) from the start as cheap hygiene that makes the eventual BAA path far shorter.

The funnel (B2B2C)

therapist hears about it (outreach / word of mouth)
  → recommends HiveJournal to clients
  → client journals
  → client shares JQ Bridge metadata with the provider (connection_type=healthcare) [CONSENT]
  → provider sees client vitals + gets engagement alerts
  → provider recommends to more clients ⟲

The provider's value (between-session visibility, a reason to re-engage a fading client) is what drives the recommend loop. The client never loses control; the provider never sees content.

Surfaces

Reuse vs build

Reuse (already there): jq_bridge_connections + jq_bridge_permissions (invite/accept/consent/revoke/metadata-toggles); trend signals (GET /api/journal/mood-stats, /api/journal/streak → days_written_30d/last_journaled_at, note_analyses.mood_trends.trend_direction); Resend email + the drip/cron patterns; the /jq-bridge + /about/jq surfaces.

Build: a provider role/account; a client-roster query + /dashboard/provider/* dashboard; a vitals snapshot endpoint (streak + last-journaled + trend-direction per client); the engagement-alert cron + provider_alerts; the provider_access_log (audit); later, license verification of professionals.

Funnel & outcome instrumentation

Acquisition funnel (mirror the Odessa funnel): provider_landing_view → provider_signup → jq_bridge_invitation_sent (exists) → jq_bridge_invitation_accepted (exists) → provider_client_viewed → provider_alert_sent. So the B2B2C funnel is measurable end-to-end.

Outcome instrumentation — the evidence layer (this is the asset; build it from day one). Per the business model below, an enterprise buyer pays for proof, not engagement, so we must capture outcomes, not just funnel events:

• Engagement-over-time per consenting client (already derivable): journaling consistency, streaks, mood-trend direction — the leading indicators.
• Validated measures, opt-in: periodic light self-report (e.g. a PHQ-9 / GAD-7 style check, consented) gives a standardized, defensible outcome variable enterprise health buyers recognize. Without a validated instrument, "they felt better" isn't purchasable evidence.
• The buyer's metric, mapped: for the disability/claims wedge that's claim duration / return-to-work; for a payer it's remission / utilization. Design the data model so a consented cohort's engagement can be linked to the buyer's own outcome metric in a pilot.
• Cohort hygiene for an eventual study: stable IDs, timestamps, consent provenance, and the provider_access_log — the difference between "we have logs" and "we have a dataset a study (or a BAA) can stand on."

Business model — evidence-first, enterprise endgame

Direction set 2026-06-13. The product is free for everyone — clients and providers alike — and stays that way. There is no consumer SKU and no client-side paywall. We considered gating the therapist-share behind a standalone subscription and explicitly rejected it: consumer willingness-to-pay for "let my provider see my mood trend" is near zero, and a paywall at the share step would add friction to the exact acquisition loop the whole thing is built on.

The real model: the asset is the evidence, and the buyer is an enterprise.

• Free maximizes adoption → adoption generates outcome data → the data is what an enterprise buyer pays for. This is land-and-expand / evidence-first, the same motion as the RCT thread. The thing we accumulate isn't revenue, it's proof.
• Awareness ≠ evidence. A big user count does not close an enterprise deal. Enterprise health buyers buy demonstrated cost reduction or a quality metric they're graded on — not engagement. So the free phase's product requirement is outcome instrumentation, not just growth (see Funnel & outcome instrumentation above).
• The enterprise-buyer ladder (easiest → hardest): digital mental-health platforms (license / acqui-hire a metadata-sharing layer) → EAPs / employer benefits → health systems → payers / HMOs. You don't aim straight at the payer; you climb the ladder, and each rung's pilot generates the evidence the next rung demands.
• First wedge — disability / claims insurers, not an HMO. Behavioral health is a top driver of long-term-disability claim duration and delayed return-to-work — the exact number a disability insurer optimizes. A consented, between-sessions signal that someone is engaging with care and trending up maps directly onto claim-duration / RTW outcomes. That makes a disability/claims insurer a more natural and nearer first enterprise buyer than a general payer.
• Warm channel (held privately): there is a warm relationship into the disability/claims insurer world that could serve as the introduction path — either as a referral/channel partner (lighter) or an acquirer/licensee (deeper). Hard gate: any commercial conversation through that channel is blocked until the IP / conflict-of-interest position is cleared with an employment lawyer (clean separation: own time/equipment, no third-party data; written acknowledgment that Throughline is separately owned IP). Details in private notes, not this repo.

Providers stay free, period. They're distribution, not revenue. A paid "practice" tier was considered as a runway bridge and rejected 2026-06-13 — charging practices would drag focus into dashboard-SaaS busywork (billing, seats, support) and dilute the only things that matter pre-deal: adoption and outcomes. The destination is the enterprise/payer contract, not subscription revenue.

Runway — fund-to-milestone (decided 2026-06-13)

No interim revenue. We fund Throughline to a defined evidence milestone, not to a profit line. The reframe that makes this safe:

• The scarce resource is attention, not cash. Throughline rides existing infra (JQ Bridge, Supabase, the app), so marginal cost ≈ founder time + near- zero infra/LLM spend. The day job is the runway — it funds life while Throughline costs almost nothing. Runway here means "can I sustain the time + keep infra near zero until the milestone," not "months of burn."
• Non-dilutive is the parallel track. Behavioral-health evidence generation is fundable (SBIR/STTR, foundation grants, academic partnerships) — and the money pays for the same study that becomes the moat. A published, grant-backed, academically co-authored result is worth far more to an insurer than the same data self-reported. Same motion as the RCT thread (Navsaria).
• Plan of record: bootstrap lean (default, affordable on day-job income) + pursue an academic co-investigator / non-dilutive grant as a high-priority background thread. Not mutually exclusive.
• The discipline this demands: "free until the milestone" rots into "free forever, no proof" without a defined, timeboxed milestone — drift, not cash, is the failure mode here. So name it.

Evidence milestone — TBD (fill in the numbers). Shape: ~N consenting clients × ~M weeks of paired engagement + validated-measure (PHQ-9/GAD-7-style) data showing a credible directional signal, mapped toward a buyer metric (disability/claims = claim duration / return-to-work). That number is what the runway funds us to, and what the first insurer pilot's evidence bar (outcome, effect size, observational vs. RCT) gets judged against.

Phased plan

• Phase 0 — design partner (no new code). Run the workflow with one real therapist (Amy Torn) + a few consenting clients on existing JQ Bridge (client invites therapist, healthcare). Learn what a therapist actually wants to see between sessions before building a dashboard for it.
• Phase 1 — provider landing + client preset + outreach. /for-therapists, the "share with my therapist" shortcut, funnel and outcome instrumentation (the evidence layer starts now), therapist outreach.
• Phase 2 — the provider product. Roster + vitals dashboard, engagement alerts, audit log. Free to providers.
• Phase 3 — evidence. Accumulate a consented cohort with validated outcome measures; stand up the dataset (and ideally a small study / RCT arm) that maps engagement to a buyer's metric. This is the real moat — no enterprise sells without it.
• Phase 4 — enterprise wedge. Take the evidence up the buyer ladder, starting with a disability/claims insurer pilot (claim-duration / return-to-work), via the warm channel — after the IP/conflict-of-interest carve-out is cleared. HIPAA/BAA + identifiable-data handling become load-bearing here. Endgame: a payer/enterprise contract or acquisition, not a subscription line.

The warm lead — Amy Torn (San Diego, CA)

A former therapist of the founder who was interested in early HiveJournal + the metadata-sharing concept years ago. The ideal Phase-0 design partner — like Navsaria for the RCT, the right first conversation. Draft outreach:

Subject: The journaling-with-your-therapist idea — finally real (and I thought of you)

Hi Amy,

Years ago you took an interest in what I was building with HiveJournal — the idea of sharing journaling insights with a provider. It's real now, and you were the first person I thought of.

Here's the shape: a client journals, and with their explicit consent shares only the trends with their therapist — mood trajectory, how consistently they're writing, recurring themes — never the actual entries. The therapist gets a little between-session visibility (and, eventually, a gentle nudge if a client goes quiet for a couple of weeks), without ever reading anything private.

I'd love your read as a clinician: would that actually be useful between sessions? What would you want to see — and what would you not want to be responsible for seeing? I can set you up to try it with a couple of consenting clients on what already works today, and I'd shape the real thing around what you tell me.

No pressure at all — even 30 minutes of your honest take would be a gift. It's good to be back in touch.

{Your name}

Status: plan for review. The Phase-0 "validate on existing JQ Bridge before building the dashboard" sequence is the considered recommendation; compliance posture (wellness-first) and business model (free-to-providers) are the live decisions.