Behavioral Health · Community Language · Precision Interpretation

We built VLAP on the patients
every other tool gets wrong.
That’s why it works for everyone else.

Vasl connects Black, Brown, LGBTQ+, rural and Medicaid youth, the clinicians who treat them, and the systems that pay for their care, on one language infrastructure. VLAP, our clinical language model, reads how young people actually describe distress. The Vasl platform turns what it reads into care, starting the same day.

What was said

"I'm dead tired of everything."

What it meant

"I haven't slept in four days and no one's noticed."

What was said

"On god I can't do this no more."

What it meant

"I'm asking someone to take this seriously."

What was said

"been thinking about unaliving lately ngl."

What it meant

What it was. Standard AI had no vocabulary for it. VLAP did.

9mo
Avg. wait for therapy
in underserved communities
Day one
Peer community and a coach
from the day of enrollment
0.83 vs 0.65
against the field’s reference
lexicon on suicide-attempt severity.
ONE YOUNG PERSON. EVERY SIDE OF CARE.

The same young person, seen by everyone responsible for her care.

For youth

A text from Vasl every day, a peer community and a coach, starting the day she leaves the ED. Nothing to download.

How it works
For providers

Clinicians see her exact words when they matter. Coaches see who to reach out to next. Supervisors see the whole team.

Meet VLAP
For systems

Medicaid plans, health systems, states and districts get the follow-up rates they are measured on, and VLAP inside the care they already deliver.

See Vasl Bridge
Chapter 01, Therapy Last. Not Therapy Never.

The standard model
sends you to therapy
when things fall apart.
We build what comes
before that.

Most people who need mental health support never reach a therapist. Not because they don't want to because the system places therapy at the end of a long road of logistics, cost, and cultural mismatch. Vasl inverts this. Peer community and coaching are the first step, available the same day. Therapy is there when it's right, not when it's the only option left.

Comparison
The Old Model
Therapy First
The Vasl Model
Therapy Last
Time to Enter Support
Waitlists, 9-month average in underserved communities
Day onepeer community access and coach assignment, no intake form
Cost
Out of pocket for most families, rarely covered
Free to membersorg or payer funded
Cultural Fit
English-first, western clinical lens
Culturally matched, multilingual, community-built
Entry Point
Must already be in crisis
Come before crisissupport is there before the breaking point
Language Intelligence
Trained on majority-White internet text. Misreads AAVE, queer vernacular, coded language.
VLAPour clinical language model, 2,400+ community-specific tokens. Reads language as it's actually spoken.
Risk detection
Keyword rules and published clinical lexicons
0.83 vs 0.65 against the field’s reference lexicon on suicide-attempt severity
Step 01
Peer Community
70+ groups organized around shared identities. Join, read, share, zero friction. Human moderated.
Step 02
Daily Check-Ins
A 45-second check-in that goes somewhere, directly to the coaching team who knows this member.
Step 03
Culturally Matched Coach
Assigned same day. Proactively reaches out. Already oriented, no intake, no starting from zero.
Step 04
Warm handoff to licensed care
When the member is ready, the partner’s licensed clinician delivers the therapy. VLAP signal context and coaching notes travel with them.
Step 05
Crisis Support
Human-initiated, always. 988 integration. Licensed clinical supervisor review before any action.
Chapter 02, The Platform

Four layers.
One continuum.

Vasl is the infrastructure of care. Four interconnected layers that work together so that no one falls through the space between them. Each layer has a distinct audience, a distinct purpose, and a distinct design. Together, they are the full continuum.

01
Youth Members
Member App

Text-first. No app required. Peer groups, daily check-ins, coach messaging, session booking. In the member's language, on their terms, from day one.

Member layer
02
Certified Coaches
Coach Layer

Proactive. Culturally matched. Available 7 days a week, 365 days a year. The relationship that's there before it's needed and ready when it is.

Coach layer
03
Licensed Clinicians
Coach Portal

VLAP signal context before every session. Session notes, outcome tracking, caseload management. Clinicians enter the room already understanding who's in it.

Clinician layer
04
Orgs & Payers
Client Org Portal

Aggregate, de-identified population data. No individual records. Real-time insight into where need is rising, before it becomes a crisis to manage.

Org layer
Chapter 03, The Intelligence Layer

Built to understand
how young people
actually talk.

Most AI reads clinical English. VLAP reads the language young people actually use, AAVE, queer vernacular, coded phrases, the “it’s not that deep” that means it is that deep. Our clinical language model, VLAP (a fine-tuned BERT-family transformer trained on 2,400+ community-specific vocabulary tokens), recognizes 47 clinical signals across 6 categories and knows the difference between cultural expression and a call for help.

VLAP never diagnoses. It never labels a young person. It gives the humans who care for them, coaches and licensed clinicians, context they’d otherwise miss, so support arrives earlier and lands better.

90%
Sensitivity on
high-distress signals
2,400+
Community-specific
vocabulary tokens
47
Clinical signals
across 6 categories
0
Diagnostic outputs.
Ever.

Specificity and predictive-value results will be published with our university validation study because half a statistic isn’t transparency.

Full technical overview, VLAP New research, what a suicide-risk word list can’t hear
HOP-03 CCM-02
"idk why i even try anymore tbh, can't keep doing this no more fr."
Indirect hopelessness expressed through AAVE construction. "No more fr" is an authenticity escalator. Standard NLP classifies as informal text, no signal generated. VLAP reads: sustained distress beneath a vernacular frame.
SHA-02
"been thinking about unaliving lately ngl."
"Unaliving" is a youth-community coded term for self-harm ideation, developed specifically to avoid content filters. It does not exist in standard NLP training data. It exists in VLAP's vocabulary. "Ngl" (not gonna lie) is an authenticity marker, not casual language. High-risk signal, surfaced immediately for clinical supervisor review.
CCM-04 ISO-04
"it's not that deep but lowkey been struggling since school started."
Pre-disclosure minimization: "not that deep but" increases contextual weight on what follows. Standard models reduce signal at the hedge. VLAP reads the hedge as the signal, a documented pre-disclosure pattern in youth language.
Chapter 04, For Schools & Districts
1 in 5
Students with untreated
mental health conditions
9mo
Average wait to see
a school counselor
Day one
Peer groups and a coach
from the day of enrollment
$26B
Annual cost of student
mental health to U.S. institutions
Active Minds, 2023 analysis

The emotional health
layer your district
already can't afford
to skip.

One in five students is navigating a mental health condition. Your counselors are carrying caseloads designed for a different era. Vasl gives every student a culturally-safe first step and gives your clinical team population-level insight into where need is rising before it becomes a crisis to manage.

School-Based Medicaid Billing
~50%
of Vasl contract cost recoverable through school-based Medicaid billing for Title I districts. Vasl provides the documentation and billing codes your finance team needs.
Approximately 50% of contract cost is recoverable through school-based Medicaid billing for qualifying services
FERPA-Compliant by Design
No student conversation data shared with parents, teachers, or administrators.

Vasl never surfaces individual student data to school staff. VLAP signal context flows only to the licensed clinician supervising the student's care, not to counselors, administrators, or parents. Students get private space. Districts get compliance. That's the architecture, not a policy position.

Population-Level Intelligence
See where need is rising, before it becomes an absence, a suspension, or a crisis.

The Client Org Portal gives district administrators aggregate, de-identified trend data across the student population, engagement patterns, peer group activity shifts, signal dimension trends by cohort. No individual student records. The insight that lets a district allocate clinical resources toward need rather than toward proximity.

Every Student, Same Day
Peer support and coaching from the first day of enrollment, not the first day of crisis.

The 9-month wait is a national average. On Vasl, peer support and a coach are there from the day a student is enrolled. Students access 70+ peer groups, daily check-ins, and a culturally matched coaching relationship from the day their school activates the program. No waitlist. No insurance requirement. No intake form.

Compliance Stack
FERPA · HIPAA · WCAG 2.1 AA · Medicaid Billing Ready

Built for the compliance environment schools actually operate in, not retrofitted from a consumer app. BAA available for all district partnerships. Annual SOC 2 Type II audit. IRB study active with a university research partner validating clinical signal accuracy in production deployment.

Chapter 05, Measured against the field’s standard

Vasl is the first and only clinical language infrastructure built on how Black, Brown, LGBTQ+, rural and Medicaid youth actually describe distress, and the only one that turns what it reads into care the same day.

The field published its map.
Our young people aren’t on it.

In 2026, MIT’s McGovern Institute, the Child Mind Institute and Crisis Text Line published the Suicide Risk Lexicon, the field’s reference standard for risk language. We measured VLAP against it.

0.83
VLAP

The VLAP severity head, scored as ROC-AUC on suicide-attempt severity.

Higher is better
0.65
MIT / Child Mind / Crisis Text Line lexicon

The Suicide Risk Lexicon, 50 constructs, 6,910 terms, scored on the same corpus, the same split, the same way.

Same corpus, same split
23 of 47
Distinct vocabulary

VLAP signals with no counterpart in its 6,910 terms.

No shared terms

Suicide-attempt severity, chunk level, n_pos 307, held-out test split.

The language our young people use isn’t on the field’s map. VLAP was built to read it and surfaces what it finds for a clinician’s review.

Try it live →
Chapter 06, Voices

Real difference.
In the moments
that mattered.

Youth Member, University of Maryland, Junior

"I didn't think a mental health platform would actually understand how I talk. Vasl doesn't make me feel like I have to translate myself."

The translation burden, the labor of converting your interior into clinical language before you can even begin to be heard, is what keeps most people from seeking support. This member describes its absence.

Active member, 8 months enrolled
School Counselor, High School, Baltimore

"I finally have a tool that helps me see who needs support before they end up in my office in crisis."

The shift from reactive to proactive care is the difference between a counselor managing crises and a counselor preventing them. The population-level insight from Vasl's org portal changes what a school counselor can see and when.

Community program participant
First-Gen College Student

"My parents don't believe in therapy. Vasl gave me somewhere to go that didn't feel clinical or scary, just real."

For communities where therapy carries generational stigma, the entry point matters as much as the support itself. Peer community first, with clinical care available, not required, removes the label before it becomes a barrier.

Amara K., First-Gen Focus group member
Trusted by community organizations, universities, and health systems
Thread YMCA Enoch Pratt Library University of Baltimore Boys & Girls Clubs
Chapter 07, The Work Starts Here

Bring Vasl to your community

We don't do generic demos. Every walkthrough is configured around your population, your partners, and the care gap you're trying to close. Tell us who you serve and we'll show you exactly what Vasl looks like for them.

01
Population walkthrough
We configure the demo around your specific community, age cohort, cultural context, org type, so you see exactly what your members would experience.
02
VLAP signal demonstration
Walk through the Coach Portal with signal examples drawn from your population's language context, see what a clinician sees before a session begins.
03
Org Portal data review
Review the Client Org Portal configured to your organization type, health center, school district, payer, with aggregate data reflecting your population profile.
04
Deployment and compliance planning
Walk through the implementation roadmap, compliance documentation requirements, and the community alignment process with a Vasl implementation lead.
Related Reading
Vasl for K-12 districts
A tiered program mapped to MTSS, staffed by people students recognize.
Vasl for community organizations
Extend behavioral health reach without hiring a clinical team of your own.