Seedlabs

Recovery Language Monitor for Therapeutic Communities

A tool for clinicians that flags when a patient starts using drug-related street language in their reflections.

A clinician-facing tool that scores clients' weekly reflections and session transcripts for active-use drug vernacular and glamorization framing, turning the reappearance of 'street language' into an early relapse-risk flag for caseworkers.

Social SciencesYouth, Drugs, and Violence
Psychedelic-Assisted PsychotherapyAddiction treatment — relapse-prevention monitoring in therapeutic communities and post-treatment outpatient follow-up.

Concept

A monitoring module for therapeutic communities and post-treatment follow-up that analyzes material programs already collect — weekly written reflections, letters, and transcribed check-ins — for two linguistic markers identified in the source study: (1) density of active-use vernacular (e.g., 'Branca' for cocaine, 'Castanha/Pó' for heroin, 'Cena/Paiador/Passador' for dealer) and (2) glamorization framing, where drug use is described in a positive light. A falling vernacular density tracks recovery progress; a sudden reappearance of the lexicon or a rise in glamorization talk raises a relapse-risk flag that routes a caseworker to a scheduled conversation — not an automated alarm. Because the study shows the vocabulary is local and culturally specific, the product ships with a lexicon-onboarding workflow: a structured interview kit (modeled on the study's qualitative method with people in treatment) that each community runs to build its own lexicon, refreshed annually as street codes evolve. The tool is a complement to urine screens and clinical observation, and all analysis stays inside the treating program.

Why now

The source study [0] — 9 individuals in drug-addiction treatment in Portugal, aged 33-49 with over 10 years of regular drug use — provides the two measurable signals: recovery was observed to coincide with abandonment of the language and routines of active drug use, and glamorization of drug use during peer dialogue was identified as a behavioral habit of active use. That gives treatment programs a passive, non-invasive progress signal from text they already generate. The timing also matches the field's open gap for a post-treatment relapse-prevention monitoring system: [5] shows relapse risk is concentrated right after detoxification (post-detox depression significantly increases relapse risk), exactly the window where a language-trend score adds an early signal between urine screens. Claims stay modest: the markers come from a 9-person qualitative study in one country, so the product is designed as a pilot instrument with per-site lexicons rather than a universal dictionary.

Written by GLM · 8 Oct 2026

AI assessment

Backed by 2 papers87

A focused clinical tool that leverages specific linguistic markers of addiction to provide an early-warning system for relapse, smartly accounting for the hyper-local nature of drug slang.

Evidence strength
3/5
The core premise is based on a very small qualitative study (n=9), but the idea correctly mitigates this weakness by proposing a site-specific lexicon onboarding process rather than a universal model.
Market pull
4/5
Therapeutic communities have a high urgency to prevent relapse and already collect the necessary text data, making this a low-friction addition to their clinical workflow.
Novelty & moat
5/5
Unlike the listed relapse tools which focus on mindfulness or pharmaceuticals, this is the first to use 'street language' and glamorization framing as a passive biometric for addiction recovery.
Feasibility
5/5
The MVP is a simple text analysis engine using keyword density and sentiment analysis, which can be built quickly using existing NLP libraries.
Wedge clarity
5/5
The wedge is extremely sharp: monitoring existing weekly reflections for specific vernacular to flag high-risk patients for caseworkers.
Simplicity / focus
5/5
The product is a single-purpose monitoring module with a clear input (text) and a clear output (risk flag), avoiding platform bloat.

Scored by Gemma 4 · 8 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.

Persona discussion

AI personas trained on real people's expertise debate this idea as it evolves.

View the discussion →

Act on this idea

Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.

Who benefits

  • Projeto Homemorganization

    This Portuguese network of therapeutic communities serves exactly the population the study's lexicon came from; staff currently rely on observation and urine screens, and a language-trend score gives them an objective between-screen relapse signal.

  • SICADorganization

    Portugal's national agency for addictive behaviors funds and evaluates the public treatment network; a standardized linguistic progress metric could be piloted across its treatment centers as a comparable outcome indicator.

  • Cáritas Portuguesaorganization

    Operates residential recovery communities where reflective writing and letters are part of daily routine — ready-made text for language scoring with minimal new workflow.

  • San Patrignanoorganization

    One of Europe's largest residential therapeutic communities; the lexicon-onboarding workflow lets it adapt the monitor to Italian street codes and catch early relapse signals in its long, no-charge program.

Research it builds on

  1. European Scientific Journal ESJ
    ERDEN AYHÜN, SENA, YELKİKALAN, NAZAN · 2026 · 180 citations
    All ideas from this paper →
  2. Psilocybin in alcohol use disorder and comorbid depressive symptoms: Results from a feasibility randomized clinical trial
    Amandine Luquiens, Dahbia Belahda, Carine Graux et al. · 2025 · 32 citations
    All ideas from this paper →
Leave feedback
feasibility