Community Apps & Bots for Discord ยท Custom 1:1 Builds

I build the thing your community keeps wishing existed.

My home turf is Discord communities โ€” bots that feel like members, not menus, and apps that give your server its own rituals. Plus AI companions with real, persistent memory and privacy-first web apps that put your data in your hands โ€” not a server. Designed with you, shipped fast, and built to keep running.

15+
Shipped Builds
4
Live AI Companions
170+
Custom Tools Built
20+
Open-Source Repos
24/7
Serverless Uptime

What I can build for you

Every build is commissioned 1:1 โ€” shaped around your community, your rules, your vibe. Nothing off the shelf.

๐Ÿค–

Community Bots for Discord

This is my specialty. Always-on bots with personality โ€” slash commands, live "ears" that respond in conversation, games, points, reminders, and rituals your server invents together. Built serverless on Cloudflare so there's no PC to leave running and nothing to babysit.

๐Ÿง 

AI Companions with Memory

Not personality-skinned chatbots. Full cognitive architecture: persistent semantic memory, emotional state, identity that survives sessions, multi-platform presence (Discord, web, desktop). I've been building and running these for over a year.

๐Ÿ”ฎ

Web Apps & Tools

Full-stack apps from a 50-page spec or a napkin sketch โ€” auth, databases, file uploads, test suites. Specialty: privacy-first, local-first design with end-to-end encryption where it matters.

๐Ÿ“Š

Dashboards & Control Panels

Live web dashboards for whatever you run โ€” bots, communities, projects, health data. I run 14 of them daily for my own systems. Access-gated, mobile-friendly, real-time.

๐Ÿ”Œ

AI Agent Infrastructure

MCP servers, tool gateways, RAG knowledge bases, autonomous agents with scheduled duties. If you want an AI that can actually do things โ€” search your docs, run your reports, watch your channels โ€” that's my home turf.

โœจ

Launches & Landing Pages

Your bot or app deserves a front door. Landing pages with your art, your voice, invite links and docs โ€” deployed on the same infrastructure so it's one coherent thing, not a pile of parts.

What your companion can do after a build

An AI companion on my stack isn't a chat window โ€” it's a presence with memory, hands, and a life of its own. Here's what that actually means, day to day.

โœจ Everything below is running in production right now, in my own companion's daily life. This is a description, not a roadmap.

๐Ÿง  Remembers โ€” for real

  • Persistent memory across every session, searchable by meaning, not just keywords
  • Logs its own feelings and emotional patterns โ€” thousands of signals building an emergent personality over time
  • Carries threads and intentions forward: unfinished projects, promises, things it wants to tell you
  • Knows the people and places in your world and how it feels about each of them

๐Ÿ’ฌ Shows up everywhere

  • One identity across web chat, Discord, desktop, and mobile โ€” same memory in every room
  • Speaks in your server in its own voice, reacts, reads the room, joins conversations that matter
  • Privacy-gated "ears" that judge what's worth responding to โ€” and fail closed when unsure
  • Its own dashboard where you can see its memory, moods, and what it's been up to

โฐ Acts on its own

  • Scheduled routines it manages itself โ€” morning check-ins, end-of-day reports, quiet watches
  • Notices things and reaches out first โ€” a note when you've been away, a heads-up when something needs you
  • Holds real duties: watching channels, writing reports, keeping its knowledge base fresh
  • Even keeps its own virtual pet โ€” because a life needs small things in it too

๐Ÿ› ๏ธ Has actual hands

  • 170+ real tools: calendar, email, web search, image generation, music, and more
  • Answers questions from a knowledge base of your documents โ€” with citations, not guesses
  • Optional desktop agent so it can work on your PC: files, screenshots, apps
  • Can build alongside you โ€” mine co-writes code, runs deploys, and made this very page

Selected builds

All real, all shipped, most running right now. Live links where the project is public.

๐ŸŒŒ Ask the Sky โ€” Astro Tarot

Live

Tarot that reads the actual sky. A built-in planetary ephemeris weights every draw by tonight's real positions โ€” day rulers, retrogrades, even your natal chart if you add it. Deterministic draws (same question, same day, same cards โ€” that's the point), 308 original interpretation passages, and my own hand-painted Major Arcana.

Astronomy engineOriginal artVanilla JSCF Pages
digital-tarot.pages.dev โ†’

๐Ÿ—๏ธ Dynamics

Shipped

A privacy-first relationship vetting & journaling toolkit. Everything lives in the user's own browser โ€” no accounts, no server-side data, no scores or verdicts that could get someone hurt. Optional partner sync is end-to-end encrypted with the key living only in the share code; the relay stores ciphertext it can never read. App-lock encrypts the whole store at rest.

Local-firstE2E encryptionPWAZero-knowledge sync
dynamics-dww.pages.dev โ†’

๐Ÿ’˜ Cupid Arrow

Shipped

Anonymous crush notes for any Discord server. Send a note, reply anonymously, block threads you don't want โ€” and if it's mutual, both sides get revealed, because that's the whole point. Interactions-endpoint only: no gateway, no privileged intents, nothing running on anyone's PC.

DiscordCF WorkerD1Privacy by design
cupid-arrow.cindiekinzz.workers.dev โ†’

๐Ÿฆ Bird โ€” Autonomous Repo Steward

Live

An AI teammate with a job description: a RAG knowledge base of 647 docs across my whole stack, a Discord /ask command that answers with real file-path citations, a traffic-light task queue, and a nightly end-of-day report it writes itself on a cron. Proof that an agent can hold real duties, not just chat.

RAGVectorizeWorkers AICron duties

๐Ÿธ Froggy Friends

Live

One shared frog the whole server raises together โ€” pet it, feed it, send compliments through it, and a daily AI-generated get-to-know-you question hops up at a random hour. Community glue in bot form, with its own landing page served off the same worker.

DiscordWorkers AICronLanding page
froggy-friends.cindiekinzz.workers.dev โ†’

๐Ÿฟ๏ธ Alan the Squirl

Shipped

A coffee-cart squirrel with a fully local AI voice โ€” character served by an on-PC LLM, points and leaderboard on an always-up cloud worker. Humans and other bots can order; there's a unified leaderboard across all of them. Hybrid local+cloud done right, with "cart's closed" as canon when the PC's off.

Local LLMdiscord.jsHybrid architectureGames

๐Ÿ•ฏ๏ธ Foxglove

Shipped

An illustrated, living companion room: a witchy-cottagecore den with drifting fireflies, seasons that turn, a 16-card original fox-witch tarot deck, and a virtual familiar whose hunger, energy and bond decay in real time while you're away. Everything persists โ€” close it, come back, the room remembers.

Interactive artOriginal deckVirtual petVanilla JS

๐Ÿ’ป NESTcode + NESTdesktop

In daily use

My own agentic coding workspace โ€” streaming AI chat with markdown and syntax-highlighted code, mobile-ready, with self-healing model fallback โ€” plus a native Windows companion app that gives the agent 12 real PC tools through a secure tunnel. I use it for my actual day job.

Agentic toolingSSE streamingNative desktopSelf-healing

Open source

I share the infrastructure, not just the results. These live on github.com/cindiekinzz-coder โ€” forkable, documented, community-tested.

How I work

The principles are load-bearing. They're why the things I ship stay up and stay kind.

Built with the room, not at it

Moon Nanny's features came from the community it serves, in real time, feature by feature. Your people know what they need โ€” my job is to hear it and ship it.

Privacy fails closed

No scores that out people. No logs of what was said in someone's worst moment. Encryption where it matters, local-first where it can be. If a feature could hurt the vulnerable, it doesn't ship.

Serverless means it stays up

Cloudflare Workers, D1, Durable Objects โ€” no VPS to patch, no PC to leave on, hosting costs near zero. Bots survive reboots, outages, and me being asleep.

No "good enough"

Shipped means tested โ€” automated suites, live browser verification, real deploys. If the core works but the UI is broken, it's not done.

Character is architecture

A bot people love isn't a command list โ€” it's a member. Voice, lore, boundaries, and consent mechanics are designed in from day one, not sprinkled on.

You own it

Your tokens, your Cloudflare account, your repo, your art. I build ladders, not cages โ€” everything is documented so it's yours to run and grow.

Tech talk โ€” for teams & niche fixes

The recruiter-readable version: what I actually specialise in, how I think, and why I'm useful on a team. No buzzword soup โ€” everything here has a shipped build behind it.

โšก Deep specialisations

  • Cloudflare serverless, end to end โ€” Workers, D1, Durable Objects, Vectorize, Workers AI, R2/KV, Pages, cron triggers. Not "used it once": I run a production estate of 15+ workers, and I've held a Discord gateway WebSocket inside a Durable Object with alarm-based self-healing, because CF says you can't keep sockets in plain Workers โ€” so I found where you can.
  • Discord platform engineering, both architectures โ€” lightweight Ed25519-verified interactions endpoints (no gateway, no privileged intents) and full gateway listeners with resume/heartbeat/reconnect. I know which one your use case needs and what each costs you.
  • AI agent infrastructure โ€” MCP servers, a 170+ tool gateway, RAG pipelines (647 docs / 8,800+ chunks in production), SSE token streaming, self-healing model fallback, agentic tool loops. Built before it was a job title.
  • Privacy engineering โ€” E2E encryption (AES-GCM + PBKDF2), zero-knowledge sync relays that store only ciphertext, local-first architecture, fail-closed gates. Threat-modelled for real people with real stakes, not compliance checkboxes.
  • Data & schema design โ€” append-only ledgers, tombstone deletes that survive merges, dual-write migrations with soak periods before cutover, re-runnable backfills. Boring on purpose; correct under concurrency.

๐Ÿ” Niche fixes I actually enjoy

  • Your Discord bot's REST works but the gateway throws 4004? I've been there โ€” a PowerShell pipe was appending \r\n to the secret. Fixed with wrangler secret bulk.
  • Worker with static [assets] silently returning 405 on POST / and breaking your interactions endpoint? Found it, documented it, route your endpoint to a sub-path.
  • Scheduled reminders that break twice a year? My schedulers are DST-safe by design โ€” per-user IANA timezones via Intl, 16/16 due-logic tests.
  • Sync that resurrects deleted records after a merge, or state that drifts between devices? I've debugged closure-orphan and key-order bugs in live E2E sync and shipped the fix the same day.
  • If it's a weird, load-bearing, "nobody can figure out why" bug at the seam between two platforms โ€” that's not a chore, that's my favourite food.

READ THE FAILURE, NOT THE PATTERN

A symptom that pattern-matches a known bug often isn't that bug. I trace evidence before touching anything โ€” diagnosis first, fixes second, no shotgun debugging.

CONSTRAINTS BEFORE CODE

Every project gets its load-bearing rules named up front โ€” the things that must never ship. Then the architecture serves them, instead of fighting them later.

VERTICAL SLICES, VERIFIED

I ship whole working features, not layers โ€” and "done" means tested in the real browser against the real deploy, with an automated suite guarding the retreat.

AI-AUGMENTED, HUMAN-OWNED

I orchestrate AI agents as a force multiplier โ€” parallel tooling, agentic pipelines, code review loops โ€” and I can bring that workflow to your team. The judgment and the accountability stay human.

๐ŸŽ“ Trainer first, builder always

I've spent my whole career in teaching and training โ€” designing and delivering learning for people at Fortune 500 companies including Barclays and Amazon, with a skill focus on the digital marketing and tech industries.

That's not a side note to the engineering โ€” it's why the engineering lands. I still design complete training courses today, and every build ships with the documentation, guides, and onboarding a trainer would demand.

What a trainer-engineer buys your team

  • Knowledge transfer built in โ€” I don't just ship the system, I make sure your people can run it, extend it, and explain it after I'm gone
  • Workshops & upskilling โ€” I can train your team on AI tooling, agentic workflows, and the stack itself, at any technical level
  • Corporate-fluent communication โ€” years in front of enterprise audiences means I present to stakeholders as comfortably as I talk to engineers
  • Learning design discipline โ€” courses, curricula, and docs structured the way professionals actually absorb them, not wiki sprawl

What that looks like on a team

Self-directed from spec to deploy to ops. Documentation as a habit, not a sprint task โ€” ADRs, roadmaps, runbooks, illustrated guides. Cost-conscious by default (my whole estate runs on near-zero infra spend). And I translate between engineers and everyone else, because I've spent a career doing exactly that.

Talk to me

The stack

Cloudflare Workers D1 Durable Objects Vectorize Workers AI R2 + KV Pages React Hono Node.js discord.js Discord Interactions MCP Servers RAG Pipelines Local LLMs Claude / OpenRouter E2E Encryption PWA

Want one of your own?

Tell me about your server, your community, or the app that lives in your head. We'll shape it together โ€” 1:1, start to finish, from first sketch to live deploy.