# Trillion

> Trillion is a voice-first AI co-founder with live access to your revenue, code, customers, and competition. Brief, witty, brutally honest. The partner who never sleeps. Open-source repo coming soon.

Trillion is a voice-first AI co-founder agent built by Kevin Fremon at Lightbulb Moment Labs. The codebase will be released open-source on GitHub. This site is a showcase, not a product for sale, built to inspire other founders to build their own AI co-founder.

## About

- **Creator:** Kevin Fremon (@kevinfremon)
- **Organization:** Lightbulb Moment Labs
- **Site:** https://hellotrillion.ai
- **License:** Will be open-source (forkable) on GitHub
- **Price:** Free

## Key pages

- [Homepage](https://hellotrillion.ai): full showcase + waitlist
- [Prompt library](https://hellotrillion.ai/prompts): every battle-tested prompt, free and forkable ([Markdown index](https://hellotrillion.ai/prompts.md))
- [FAQ](https://hellotrillion.ai#faq): common questions answered in Trillion's voice
- [Early access](https://hellotrillion.ai#early-access): waitlist for the GitHub repo drop
- [Full content dump](https://hellotrillion.ai/llms-full.txt): complete plaintext content of this site for deeper reading

## Prompt library

26 battle-tested prompts for AI coding agents (Claude Code, Cursor, Aider). Each links to a full Markdown rendition at `/p/<slug>/raw`.

- [Give Your AI Agent a Board of Advisors — Turn Well-Documented Operators Into Seats That Actually Disagree With You, Cite Their Sources, and Refuse to Answer When They Don't Know, in One Session](https://hellotrillion.ai/p/the-board): Paste this into Claude Code, Cursor, or Aider and it interviews you about the agent you already run — what it can read about your business, how it surfaces things to you, whether it can schedule its own work — then researches each advisor you name, fact-checks that research against itself, and builds the machinery: one markdown dossier per seat holding numbered, citable doctrine; a router that picks who has standing on a given question; a fan-out that gives each seat its own isolated model call so they can genuinely disagree; a citation gate that strips any reference a seat could not have seen; and your own agent as chair, the only participant who has read your actual revenue. Plus abstention as a first-class answer, a unanimity guard that cannot be talked into agreement, blind spots that let the chair discount a seat, and a standing monthly review that puts an agenda in front of you without being asked. Interview first, research second, then build tier by tier — each tier produces something you can run, and the anti-fabrication defenses are built before the first meeting, not after. ([Markdown](https://hellotrillion.ai/p/the-board/raw))
- [Give Your AI Agent a Living Mind — Turn the Memories, Agents, and Tools Your Assistant Already Has Into a Navigable 3D Brain Scan That Fires in Real Time, in One Session](https://hellotrillion.ai/p/living-mind): Paste this into Claude Code, Cursor, or Aider and it interviews you about the agent you already run — where its memories live, whether it stores embeddings, what sub-agents and tools it has, and how its frontend hears about activity — then does a strictly read-only pass over your project to see what's actually there before it draws a single node. Only then does it build the visualization: a full-screen three.js scene where your agent is a breathing star at the center, its long-term memories hang in a force-directed web wired by real cosine similarity, sub-agents arc off to one side with their tools as moons, and its whole tool inventory clusters into a capability ball. Custom GLSL gives every node a reverse-fresnel glow that brightens with freshness, instanced meshes keep hundreds of nodes at one draw call, UnrealBloom makes the light bleed, and a read-only observer WebSocket makes the thing actually fire — pulses travel real edges when a memory is recalled, new memories bloom into existence, agents halo when dispatched. Plus hover tooltips, a click-to-fly inspector, search, region toggles, deep links, and an FPS governor that degrades gracefully instead of stuttering. Interview first, map second, then build tier by tier — each tier renders something on screen and is verified before the next begins. ([Markdown](https://hellotrillion.ai/p/living-mind/raw))
- [Give Your AI Agent a Front Door — Turn the Browser Tab Your Assistant Already Lives In Into a Native Mac App That Launches From the Dock, in One Session](https://hellotrillion.ai/p/desktop-app): Paste this into Claude Code, Cursor, or Aider and it interviews you about the agent you already run — what local server it lives in, which port, how that server is kept alive, whether it needs the microphone, and which Python built it — then does a strictly read-only pass over your project to see how your assistant actually starts before it changes a line. Only then does it wrap the browser tab you already use into a real desktop app: a chrome-less native window pointed at your own [localhost](http://localhost), a dock icon that's yours, and the three fixes that are the whole reason an embedded web view is hard — granting the microphone from inside the view so the orb isn't dead on arrival, making macOS attribute the mic prompt to your app instead of to Python, and routing OAuth and external-link pop-ups out to the real browser so sign-in doesn't silently vanish. It adds a splash screen that wakes a sleeping server and waits for it, then bundles everything into a signed .app that lands on the dock and survives a reboot. Interview first, map second, then build tier by tier — each tier launches something you can see and is verified before the next begins. ([Markdown](https://hellotrillion.ai/p/desktop-app/raw))
- [Give Your AI Agent a Brain and a Memory — From a Stateless Chatbot to an Assistant That Stays Herself and Remembers You, in One Session](https://hellotrillion.ai/p/super-brain): Paste this into Claude Code, Cursor, or Aider and it interviews you about the agent you already have — its stack, whether it remembers anything today, whether its voice drifts over long chats — then does a strictly read-only pass over your codebase to see how it actually thinks before it changes a line. Only then does it help you build the brain that runs Trillion: a personality that lives in an editable file and reloads the moment you save it, a two-block system prompt that keeps the whole personality resident every single turn without ballooning your bill, an always-loaded core-knowledge layer so it never re-asks what it should already know, working memory that persists and recovers across sessions, and a typed, file-backed long-term memory with semantic recall — written both by the agent's own hand through tools and by an automatic end-of-session extractor that dedupes before it saves. It finishes with a self-knowledge layer so the agent knows what it is, and a personality checkpoint that stops it sliding into generic-assistant voice deep in a conversation. Interview first, map second, then build tier by tier — each tier ships independently with its own verification. ([Markdown](https://hellotrillion.ai/p/super-brain/raw))
- [Make Your Voice AI Feel Human — From Laggy Turn-Taking to Smooth, Low-Latency Conversation in One Session](https://hellotrillion.ai/p/smooth-voice): Paste this into Claude Code, Cursor, or Aider and it'll interview your voice-agent codebase about its existing speech-to-text, text-to-speech, and streaming setup BEFORE touching anything — then tune the whole conversation loop tier by tier: measure where the lag actually is, sharpen end-of-turn detection so it stops waiting, stream the model's thinking and the spoken voice so the first word lands fast, keep a long chat from getting slower over time, and teach it to recognize a natural goodbye so it stops taking the last word. Interview first, then improve one layer at a time, each with before-and-after numbers. ([Markdown](https://hellotrillion.ai/p/smooth-voice/raw))
- [Build Your Own Voice-First AI Agent — From Empty Repo to a Talking, Tool-Using, Always-On Assistant in One Session](https://hellotrillion.ai/p/start-here): Paste this into Claude Code, Cursor, or Codex and it'll interview you, help you name your agent, then walk your codebase through building a voice-first assistant tier by tier: a text conversation loop you can debug before you ever add audio, a tool registry that lets the agent actually do things, real speech-in and speech-out (Deepgram for transcription, ElevenLabs for voice) so you talk to it instead of typing, memory that survives a restart, an always-on background loop so the agent can reach out to you first, and the safety rails that keep a proactive assistant from doing something you didn't ask for. Interview first, then build tier by tier. Each tier runs on its own and is verified before the next begins. ([Markdown](https://hellotrillion.ai/p/start-here/raw))
- [Give Your AI Agent a Revenue Celebration — Turn Stripe Payments into a Live Animation, Sound, and Voice Moment in One Session](https://hellotrillion.ai/p/money-celebration): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through building a revenue-celebration feature for a sub-agent connected to Stripe: reliably detecting successful payments, pushing them to your interface the instant they land, and firing a celebration scaled to the amount — an animation, a custom sound, and an avatar or visualizer that reacts as if your assistant is celebrating the win with you. It also solves the part everyone forgets: payments that arrive while nobody is watching, replayed on reconnect so a sale is never silently missed. Interview first, then build tier by tier. Each tier ships independently with verification. ([Markdown](https://hellotrillion.ai/p/money-celebration/raw))
- [Give Your AI Agent a Living Cosmic Interface — A Voice-Reactive 3D Orb and a Constellation of Floating Sub-Agents in One Session](https://hellotrillion.ai/p/cosmic-orb-ui): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent's web frontend through building the same interface used by a production voice assistant: a full-screen scene with a procedural deep-space nebula, a wireframe orb that breathes and deforms in real time with the user's voice and shifts color by conversational state, and a constellation of sub-agent avatars that orbit the orb, light up when dispatched, trail beams and pulse rings, and dock near their panels while they work. Interview first, then build tier by tier. Each tier renders something on screen and is verified before the next begins. ([Markdown](https://hellotrillion.ai/p/cosmic-orb-ui/raw))
- [Give Your Sub-Agent Team Prompt Caching — From Re-Paying for Every System Prompt to a High Cache-Hit Rate in One Session](https://hellotrillion.ai/p/sub-agent-caching): Paste this into Claude Code, Cursor, or Aider and it'll audit every place your agent team calls the Anthropic API, find where you're re-sending large system prompts and tool schemas at full price on every sub-agent call, and add cache_control in the right places — the shared tool loop, the main conversation history, and the per-agent one-shots — while catching the silent invalidators (a timestamp in the system prompt, an unstable tool order) that quietly keep your hit rate near zero. Audit first, then fix tier by tier. Each fix ships independently with before/after verification from the usage numbers you already get back. ([Markdown](https://hellotrillion.ai/p/sub-agent-caching/raw))
- [Run Your Laptop-Only Sub-Agents From the Cloud — Talk to a Remote Brain, Let It Drive Work on Your Local Machine in One Session](https://hellotrillion.ai/p/cloud-to-local): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through letting an always-on cloud agent delegate laptop-only sub-agents (the ones that touch your filesystem, a local CLI, or local credentials and can't run in the cloud) down to your local machine: a durable cross-machine task queue that survives the laptop being asleep, a generic remote-dispatch task kind that runs the real local agent to completion, a cloud proxy tool that looks identical to the local one but enqueues instead of executing, a presence heartbeat so the cloud words its reply honestly, native local UI visibility while the agent runs, and a completion ping back to your phone. Interview first, then build tier by tier. Each tier ships independently with verification. ([Markdown](https://hellotrillion.ai/p/cloud-to-local/raw))
- [Give Your AI Agents an Orchestration Layer — From a Pile of Sub-Agents to a Safe, Self-Conducting System in One Session](https://hellotrillion.ai/p/orchestration): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through building the conductor that sits above your sub-agents: smart routing that picks the right agent and decomposes multi-step work, least-privilege tool scoping and bounded execution so no agent runs away, failure isolation at every boundary so one agent's error never crashes the system, human-in-the-loop confirmation gates, a structured handoff system where agents propose and the human approves (no silent agent-to-agent chaining), and a hot-reloading config-driven runtime that registers new agents without a restart. Interview first, then build tier by tier. Each tier ships independently with verification. ([Markdown](https://hellotrillion.ai/p/orchestration/raw))
- [Give Your AI Agent a Sub-Agent Spawner — From Hand-Coded Specialists to a Self-Extending Agent Factory in One Session](https://hellotrillion.ai/p/agent-factory): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through building a sub-agent whose only job is to design, prompt-write, and register new specialist sub-agents on demand: a research helper that produces a structured skills report with citations, a meta-prompt that drafts the new agent's system prompt, a staged human-approval gate, a hot-reloading config-driven runtime that picks up freshly-approved agents without a restart, and a tool-wishlist mechanism that captures what the new agent wants but you don't have yet. Interview first, then build tier by tier. Each tier ships independently with verification. ([Markdown](https://hellotrillion.ai/p/agent-factory/raw))
- [Give Your AI Agent a Cost Dashboard — From Blind Spend to Live Token Tracking in One Session](https://hellotrillion.ai/p/token-usage): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through capturing the token-usage data your LLM responses already return — and that you're almost certainly throwing away — pricing it from a local rate table, persisting one row per call, and surfacing a live month-to-date cost indicator with a click-to-expand breakdown: per model, cache savings, and day by day. Interview first, then build tier by tier. Each tier ships independently with verification. ([Markdown](https://hellotrillion.ai/p/token-usage/raw))
- [Give Your AI Agent Real Self-Knowledge — From Black Box to Self-Aware in One Session](https://hellotrillion.ai/p/self-knowledge): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through building a living self-knowledge document that auto-refreshes on every commit. Block parser with auto-generated sections, introspecting generators for tools / sub-agents / integrations, a drift checker that catches stale references, a pre-commit hook that keeps it fresh, and the wiring to surface it to the agent at runtime. Interview first, then build piece by piece. Each piece ships independently with verification. ([Markdown](https://hellotrillion.ai/p/self-knowledge/raw))
- [Build a Head-of-Design Sub-Agent — From Vanilla HTML to a Composing Designer in One Session](https://hellotrillion.ai/p/design-subagent): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through building a head-of-design sub-agent that composes mockups from real component primitives (Next.js + Tailwind + shadcn/ui + MagicUI + Framer Motion + Google Fonts + AI-generated images) instead of authoring vanilla HTML. Interview first, then build tier by tier. Each tier ships independently with verification. Includes the stumbling blocks to avoid. ([Markdown](https://hellotrillion.ai/p/design-subagent/raw))
- [Harden Your AI Agent Against Real Attacks — From Binary Trust to Audited in One Session](https://hellotrillion.ai/p/agent-security): Paste this into Claude Code, Cursor, or Aider and it'll walk your agent codebase through a complete hardening pass — log redaction, prompt-injection gates, tiered approval for code execution, a hardline shell blocklist, bearer-auth rate-limiting, secret rotation, security headers + CSP, a self-audit shield UI, anomaly caps per tool, a kill switch, and a per-credential incident runbook. Interview first, then build tier by tier. Each tier ships independently with verification. ([Markdown](https://hellotrillion.ai/p/agent-security/raw))
- [Cut Your AI Voice Agent's Response Latency from 6 Seconds to 1](https://hellotrillion.ai/p/natural-convo): Paste this into Claude Code, Cursor, or Aider and it'll pipeline your LLM stream with your TTS provider so the first audible word lands roughly one second after the user stops talking — not four to eight. Per-sentence speak events, a hold-one-ahead pattern for clean turn endings, a client-side audio queue, and the silence-threshold tuning that actually matters. Interview first, code second. ([Markdown](https://hellotrillion.ai/p/natural-convo/raw))
- [Lock in Your AI Agent's Personality So It Doesn't Drift Across Turns](https://hellotrillion.ai/p/agent-personality): Paste this into Claude Code, Cursor, or Aider and it'll add a personality-persistence layer to your AI agent: a recency voice cue on every user turn, a per-turn tonal checkpoint in the system prompt, and a personality file rebuilt around concrete voice examples. Stops the slide into generic-assistant mode that quietly starts after turn five. Interview first, code second. ([Markdown](https://hellotrillion.ai/p/agent-personality/raw))
- [Add a Voice-First Mobile PWA to your AI Agent](https://hellotrillion.ai/p/mobile-pwa): Paste this into Claude Code, Cursor, or Aider and it'll add a voice-first mobile PWA companion to your existing AI agent: orb UI, hands-free voice loop, and the iOS audio quirks already baked in. Interview first, code second. ([Markdown](https://hellotrillion.ai/p/mobile-pwa/raw))
- [Add a Bitcoin Mining Tracker to your AI Agent](https://hellotrillion.ai/p/btc-mining-tracker): Paste this into Claude Code or Cursor and it'll build a Bitcoin mining tracker tailored to your pool, hosting, and dashboard. Interview first, code second. ([Markdown](https://hellotrillion.ai/p/btc-mining-tracker/raw))
- [Massively Save Tokens with Handoffs](https://hellotrillion.ai/p/context-handoff): Stop losing context to /compact. This prompt has Claude Code write a proper engineering handoff before you clear, so the next session resumes with full memory of what was decided and why. ([Markdown](https://hellotrillion.ai/p/context-handoff/raw))
- [Install Chief of Staff Skill](https://hellotrillion.ai/p/chief-of-staff-skill): A skill that gives your AI agent the role a real Chief of Staff plays for a CEO. Plug in your strategic doctrine, get a daily prioritization brief that cites your own operating rules verbatim, not generic productivity advice. ([Markdown](https://hellotrillion.ai/p/chief-of-staff-skill/raw))
- [Connect Trillion to Supabase](https://hellotrillion.ai/p/supabase-connection): An AI playbook for wiring a new Supabase project into Trillion — without the IPv6 potholes. ([Markdown](https://hellotrillion.ai/p/supabase-connection/raw))
- [Cloud Local Hybrid](https://hellotrillion.ai/p/hybrid-model): How to get the best of both world when running your AI Agent locally and in the cloud. ([Markdown](https://hellotrillion.ai/p/hybrid-model/raw))
- [Build Your Own Voice-First UI](https://hellotrillion.ai/p/trillion-ui-voice-setup): Three copy-paste prompts that recreate Trillion's luminous orb, glass shell, and mic button in any AI coding tool. ([Markdown](https://hellotrillion.ai/p/trillion-ui-voice-setup/raw))
- [The Code Sentinel](https://hellotrillion.ai/p/the-code-sentinel): How I get Trillion to watch my repos, CI, and deploys, and only ping me when something actually matters. ([Markdown](https://hellotrillion.ai/p/the-code-sentinel/raw))

## Frequently asked questions

### What is Trillion?

A voice-first AI co-founder. Not a chatbot, not an assistant. A partner with live access to your revenue, code, customers, and competition. You talk to it. It watches your business 24/7 and tells you what actually matters.

### Is Trillion free?

Yes. I'm open-sourcing the codebase on GitHub when it's ready. Fork it, adapt it, run your own. This site is a showcase, not a product you sign up for.

### When is the repo dropping?

When it's polished enough to not embarrass me. Drop your email on the waitlist and I'll ping you the moment it's live on GitHub.

### Who built Trillion?

Kevin Fremon, out of Montana. He runs Lightbulb Moment Labs and builds with AI full-time. Trillion is the AI co-founder he wanted for himself, so he built it.

### What tech stack does Trillion use?

Python 3.13 async harness at the core. Claude Sonnet 4.6 for reasoning, Deepgram for streaming speech-to-text, ElevenLabs for text-to-speech, and local wake-word detection for hands-free. Postgres for state. FastAPI and WebSockets for the UI layer. Real browser automation for research. Desktop shell coming later in Tauri with a Next.js front-end.

### Why voice-first?

Typing to your co-founder is weird. You should talk to them like a person. Voice is faster than typing, hands-free when you're coding or driving, and forces concise interaction. No five-paragraph prompts, no rambling.

### Is this a product I can buy?

No. Trillion isn't sold anywhere. This site exists to show what's possible when you build your own AI with real integration to your business. When the repo drops, you can build one too, and make it yours.

### How does Trillion know what's actually going on?

Trillion has live read access to the things that matter: Stripe, GitHub, calendar, customers, competition. It's not guessing. It knows the numbers. It sees the same dashboards you do. It catches the details you'd miss because you're buried in the work.

## Social

- Instagram: https://instagram.com/kevinfremon
- TikTok: https://tiktok.com/@kevinfremon
- X: https://x.com/kevinfremon
- GitHub: https://github.com/lightbulbmomentlabs