AI Mediator | Not Hacker News!

Discussion (1 comments)

Showing 5 comments

27 days ago

1 reply

Hello, Hacker News. I made a thing. Mostly because I was tired of watching people (myself included) detonate perfectly good relationships with a single poorly worded text message. The app is called Mitigate, and it’s basically an AI conflict mediator for humans who are trying their best but occasionally communicate like malfunctioning toasters. You can talk to it, type to it, or upload screenshots of That One Conversation™ you wish you’d handled differently. It reads tone, detects blame spirals, spots gaslighting patterns, and rewrites messages in a way that won’t turn the next three hours of your life into emotional dodgeball. I didn’t build this because I’m a genius. I built it because healthy communication is supposedly “common sense,” yet reaching for common sense when you’re angry, hurt, or jealous is like expecting a crashing server to gracefully run a patch. Meanwhile, we’re all out here raw-dogging relationships with instincts forged in the Paleolithic era. At some point in adulthood, I realized something quietly horrifying: that good communication is one of the most valuable skills a person can possess and yet, it’s never actually taught to us. We learn calculus, the mitochondria, that Christopher Columbus was a good guy (mf), but not how to say, “Hey, that hurt my feelings,” without sounding like a four-year-old. So, we keep stumbling through relationships, friendships, and group chats like well-intentioned wrecking balls, accidentally launching conflicts we never meant to start. So, I wondered: What if AI could act as a neutral third party? What if it could keep two people from accidentally becoming supervillains in each other’s origin stories? And then, after watching one too many conversations burst into flames over a misplaced tone or poorly chosen word, I finally snapped and built an AI Mediator. Under the hood, it blends structured communication frameworks with emotional-reasoning layers and a neutrality engine so stubbornly unbiased it won’t pick a favorite even if you try to bribe it. It writes guidance and scripts informed by restorative communication theory, conflict psychology, and the sincere hope that none of us truly want to die on the hill of “you used the wrong emoji.” I also spent an unreasonable amount of time refining tone detection so it can differentiate between honest confusion, weaponized ambiguity, and whatever feral emotional energy is happening in a message that says, “we’re fine .” What it actually does: • Analyzes both sides of a conflict without taking sides • Spots emotional tone shifts, blame patterns, and manipulative phrasing (intentional or not) • Gives real-time de-escalation guidance • Rewrites messages into “human but calm” instead of “human but on fire” • Lets you mediate with yourself or with another person using AI as the referee who never gets tired or snarky So, here we are. I built the anti-drama enforcer. Live demo goes up in a couple weeks. I’ll update this thread the moment it’s live. If you’re curious and want to kick the tires early (or help break things), the waitlist is open. I also added a small early-access tier for anyone who wants to explore the earliest build. I would genuinely love your feedback, critiques, warnings, philosophical rants, performance complaints, and “you fool, you should’ve done X instead of Y” comments. Be gentle. Or don’t. I made an app for conflict. I can handle it.

delichon

27 days ago

Give two leading spaces to items in bullet lists to indent them properly.

qwert12345887

27 days ago

1 reply

I was just about to start building something like this.

  * Ability to branch off—similar to Zoom breakout rooms—where you preserve context but temporarily go into a 1:1 space with the AI. Great for brief deep-dive moments.
  * Typing slows down thinking and breaks flow. Voice APIs are finally good enough that voice should be first-class.
  * Privacy matters a lot. Especially for personal sessions. The ability to completely wipe everything is critical—one preserved context across threads is enough to lose trust.

It would also be useful if both sides could select the "AI mode" for a new thread:

  * Nurture mode: healthy listening, feelings, emotional context.
  * Finance mode: results-oriented, financially logical, grounded in reality.
  * Career mode: guidance, planning, and professional reasoning.
  * Adventurous mode: creative, exploratory, high-novelty thinking and so on.

Next step: gentle preferences nudges. Maybe ask one simple question per week to learn about the user’s likes/dislikes. Make it editable and transparent.

qwert12345887

27 days ago

Also async, i just don't think real time resolution is a good thing for personal stuff, its too charged and emotional

mksinclair

27 days ago

Thanks everyone for taking a look — happy to answer anything about the build, the approach, or the decision-making behind it.

A couple notes up front: Mitigate is currently in pre-launch. The live demo goes up in a couple weeks — I’ll update this thread the moment it’s ready.

If anyone wants to try it early (or help break it), I opened a waitlist on the site. There are also a few early-access spots for people who want to explore the earliest working version. Tech details: Right now the backend is a mix of structured communication modeling, sentiment/tone classification layers, and an LLM-driven mediation pipeline. The hardest parts so far have been:

• keeping the AI strictly neutral (it loves to “fix” the wrong person) • calibrating tone detection so it doesn’t overreact or underreact • deciding when to rewrite vs. when to educate • preventing “therapy drift” — this isn’t meant to replace mental health professionals Current limitations: It sometimes errs on the side of too gentle, and nuance in long chains of messages is still tricky. Multi-person conversations (family, teams, etc.) are on the roadmap but not stable yet. What I’d love feedback on: • Are the core assumptions here reasonable • Where could the mediation logic be improved • Any obvious product traps I’m not seeing • Ethical/safety concerns around real-time mediation • Features you wish existed in conflict-resolution tools?

I’m here and will respond as quickly as I can. Appreciate the time and the critique — genuinely.

Resources