The LLM doesn't try to be the help. It routes the right signal to the right person at the right time.
A demonstration prototype, deployed today on company-owned production infrastructure. Portable to Azure Government Cloud upon Phase I award. Built around a load-bearing premise: most of the highest-risk veterans never engage the formal crisis-care system, but they do maintain peer contact through informal networks.
Every component has a direct architectural peer in Azure. The Phase I lift is a deploy-config change, not a re-architecture. The crisis flag → peer-first → VCL fallback sequence is a single deterministic state machine — independent of the LLM model selection.
Mandatory human-in-the-loop. No LLM output bypasses the deterministic escalation rule.
Every classification ships with C-SSRS / Joiner ITS framework citation and reasoning.
Escalation rule is data, not a code path inferred from model output.
No login. No identity binding. Whatever you write is hashed before anything is written down. The only thing that survives is its length and the system's read on the risk.
in · hold · out · resonance 0.1 Hz
none — framework: Columbia C-SSRS step 1 · indicators: none observed · next action: continue check-in cadence.
Phase I demonstration prototype.
Not for actual crisis support. Not authorized for clinical use. Not a substitute for licensed mental health care. Processes no patient data.
If you or a veteran you know is in crisis: call 988 then press 1, or text 838255.