Operator workspace
Provide a human-led place to select a conversation and prepare an operator response.
Clinical demonstrator
Local product demonstratorMeridian Crisis Text
A synthetic text-conversation workspace separating operator replies, consented AI simulation and Guardian oversight.
Human-led text operations. Visible oversight.
Crisis Text explores the operations around a sensitive text service: a visible conversation queue, explicit ownership, human response and a separate oversight view. The current page is a local prototype, not a public support service or a live SMS channel.
What the AI does. What people decide.
The actual assistance
AI features are controlled by an administrator setting and remain simulator-only. The interface separates human-operator mode, one consented simulated AI turn at a time and Guardian review. Switching the view does not change conversation ownership. Oversight can pause AI turns for human attention; no unattended crisis intervention is established.
Human ownership
A local prototype with an observed queue-rendering issue. Live services, practitioner verification and operational crisis-service readiness remain outside the current demonstration.
Features in the current workspace.
Conversation ownership
Keep assignment separate from the currently selected UI mode.
Queue views
Offer open, unassigned, mine, Guardian-cue, handover and closed views, plus owner filtering controls.
Simulator inputs
Load an example queue or simulate an incoming text rather than connect a live SMS gateway.
Separate AI control
Keep AI enablement behind an admin setting and simulated turns explicitly consented.
Guardian oversight
Provide an oversight mode for language cues, handoff review and pausing simulated AI activity.
Continuity context
Provide fields for a human review, safety-plan documentation and handover in the selected local case.
Operational wallboard
Expose a dedicated local wallboard view for service-operations demonstration.
From first input to a reviewed next step.
- 01
Choose a fictional case
Use a simulated conversation only.
- 02
Assign a person
Confirm the owner independently of the view mode.
- 03
Work the conversation
Use human-led responses or permitted simulation.
- 04
Review oversight
Inspect advisory cues and handoff needs.
- 05
Record continuity
Retain the intended local review and handover context.
How the product works.
| Area | Mechanism and scope |
|---|---|
| Channel boundary | The SMS gateway is disconnected and the simulator sends no external messages. |
| Identity boundary | Local demonstration access does not verify practitioner identity or qualifications. |
| AI operation | The UI distinguishes operator, simulated agent and oversight modes rather than implying a fully autonomous responder. |
| Current inspection limit | An owner-filter rendering error was observed in this local build. End-to-end queue operation is not asserted as verified. |
Product evidence and source notes
Content checked against the v25.2.0 source and visible local workspace on 30 October 2026. This is a product description and interface inspection, not an end-to-end deployment or clinical validation.
static/crisis-text.js; routes/crisis_text.py; port-8050 workspace inspection
Review assurance scopeQuestions about Crisis Text.
Is this a live crisis support service?
No. This website describes a software demonstrator and does not offer a crisis communication service.
Does it send real texts?
No external SMS messages are sent by the reviewed local simulator.