ReplyNest

A persistent community-care Mind for creators. Built for the Creative Minds Jam #1 — Moderation & community assistance track.

ReplyNest uses a Minds by Animoca Brands agent as its core decision-maker. It preserves a single conversation alias across requests, so the Mind can retain member context, resume unresolved care, and provide proactive daily follow-ups.

The built-in showcase uses synthetic creator-community records. The UI labels every Mind output as a creator-reviewable proposal; it never claims that a Mind knows a person or took an external action without real source data and an approved integration.

Why it fits the brief

Jam requirement	ReplyNest implementation
Persistent Mind agent	All decisions go through a configured Mind with ensureConversation(alias, mindId).
Memory	The Mind sees historical context through the same stable conversation, while the UI makes remembered member context visible.
Continuity	A member's unresolved issue is included in the next care decision, even after a new browser session.
Autonomous follow-up	/api/autopilot sends the care queue to the Mind; vercel.json schedules it every day at 09:00 UTC.
Creator-economy problem	Creators can care for a growing community without treating loyal members like anonymous tickets.

Run locally

Requires Node 22+.

npm install
cp .env.example .env.local
npm run dev

The interface works in clearly-labelled demo mode before credentials are added. To make every draft and follow-up come from a live Mind:

1. Create a Mind and a Builder API key from the Minds Builder account setup.
2. Add MINDS_BUILDER_API_KEY and MINDS_MIND_ID to .env.local.
3. Restart the development server.

Never put the Builder API key in a NEXT_PUBLIC_* variable or the browser: it is used only by server routes.

For deployment, add a random CRON_SECRET to your host's environment variables. Vercel uses it to authenticate the scheduled GET /api/autopilot invocation, preventing arbitrary requests from spending the Mind's Cognition Credits.

Suggested Mind DNA

When creating the Mind, tell the Concierge:

You are Echo, ReplyNest's community-care agent for independent creators. You remember community members as people, prioritize unresolved harm and access issues, write warm and concise replies, and proactively surface one meaningful follow-up each day. Never promise an action you cannot execute, never expose private member information, and ask for creator approval before any public reply is sent.

Demo script (1.5–2 min)

1. Start on the dashboard: introduce Echo as a community Mind that is always on.
2. Select Maya: show the visible remembered context, then choose Draft with Echo.
3. Refresh the page or select Leo: explain that Echo retains the same conversation alias and can resume the broken-link issue.
4. Click Show persistent memory proof: show the live Minds conversation trail — not a mocked activity log.
5. Click Run daily follow-up: show Echo prioritizing Leo without another user prompt.
6. Close on the three proof points: long-term memory, cross-session continuity, and autonomous daily follow-up.

Demo video

demo/ReplyNest-Demo.mp4 is a 99.9-second, 720p English-narrated walkthrough of the product and its Minds integration.

Submission checklist

• Register for Creative Minds Jam #1 on DoraHacks
• Add a live Mind + Builder API key
• Record the demo after verifying a live reply
• Push this repository to GitHub
• Add the public repository and 1.5–2 minute video to the DoraHacks BUIDL submission