How to Use LM Mediator: Listen, Ask, Reflect, and Build a Peaceful Next Step

You will finish with an editable transcript, optional live mediator answers, a neutral AI-assisted observation, practical next steps, and a saved final agreement without presenting the software as therapy, counseling, diagnosis, legal advice, or emergency care.

Before you start

LM Mediator is for ordinary communication and mediation-style reflection. Everyone present should know the conversation is being transcribed and agree to it. Do not use mediation to negotiate immediate danger, abuse, coercion, violence, or a self-harm crisis; put safety and qualified human support first.

Step 1: Choose the conversation setting

Choose Home, Friends, or Volunteers so the reflection has basic context without requiring a long intake form.

Step 2: Choose and preview the mediator voice

Choose Samantha, Tessa, Rishi, or Moira. LM Mediator does not silently swap in another system voice if the selected voice is unavailable.

Step 3: Get clear consent before listening

Confirm that everyone present knows transcription is happening and agrees. LM Mediator does not save an audio recording as part of this workflow.

Step 4: Listen and ask the mediator when useful

Use compatible live speech recognition or type/paste the conversation. With Conversation Mode on, start a direct question with “Mediator…” or “LM Mediator…” and LM will pause listening, answer briefly and neutrally, speak the answer, add it to the transcript, then resume listening.

Step 5: Review the transcript

Transcript text autosaves with the private LMOS project. Correct important names, facts, missing phrases, or recognition mistakes before asking for the fuller mediation reflection.

Step 6: Create the mediation reflection

LM Mediator looks for observable statements, shared goals, misunderstandings and practical agreements without diagnosing people or deciding who won.

Step 7: Read, agree and follow through

Hear the observation aloud when appropriate, work through the next-step checklist, write the agreement in your own words and save it as the final agreement in LMOS.

Make the workflow better

• Keep the transcript focused on the actual conversation rather than building a case against the other person.

• Use direct mediator questions for short clarification, not to make the AI pick a winner.

• Correct factual transcription mistakes before mediation.

• Choose one or two realistic next actions over a long list of promises.

• If someone feels unsafe, stop ordinary mediation and get appropriate human help.

Questions people ask

Is LM Mediator counseling or therapy?

No. It is a communication aid for basic mediation-style reflection and practical follow-through. It does not diagnose, treat, counsel, or replace a qualified professional.

Can LM Mediator join the conversation?

Yes, when Conversation Mode is on and someone directly addresses “Mediator…” or “LM Mediator…” with a question. The answer is intentionally brief, neutral, spoken aloud when safe, and appended to the transcript before listening resumes.

Does LM Mediator record the room?

The current workflow does not save an audio recording. Compatible device or browser speech recognition converts speech to editable text, which autosaves with the private LMOS project; relevant transcript text is sent to the AI service only for a direct mediator question or the requested mediation reflection.

What will LM Mediator cost?

The planned paid launch is one simple plan at $5 per month or $50 per year. The Founding Beta remains free while the live conversation loop is validated.

What happens if the conversation includes threats or abuse?

The mediator is designed to stop ordinary compromise framing, flag the safety concern, disable automatic read-back of the sensitive result, and prioritize appropriate human or emergency/crisis support.