-
Notifications
You must be signed in to change notification settings - Fork 539
Company Handbook
This is Hyprnote's public company handbook. We believe in transparency and sharing how we think, work, and build.
Hyprnote is an open-source AI notetaker built for people who believe they deserve bespoke software and complete control over their AI.
Beyond Note-Taking: Note-taking reduces cognitive load. We're extending that idea: what other cognitive burdens can we offload? Today it's notes. Tomorrow, email, calendar, and broader workflow - an AI-native executive assistant that actually works for you.
Your AI, Your Way:
- Use on-device models for complete privacy
- Bring any AI model - cloud or local
- Own your data completely
- No vendor lock-in
Like Apple Notes, but built for meetings. Like an assistant that records, understands, and organizes - on your terms.
Hyprnote was born from frustration.
John's sales background meant living through countless meetings where existing tools transcribed but didn't go further. As a heavy note-taker for over 20 years who has tried every app in the market - Obsidian, Logseq, Apple Notes, and more - he has very specific taste and strong opinions about what a note-taking tool should be.
Yujong's open-source background brought the conviction that solutions must be transparent and community-driven. Create massive value first, capture some of it later.
Together, we wanted craft and simplicity. Trustworthy, reliable, uncompromising.
LLMs changed everything. Ideas prototype faster, models run locally, hardware keeps improving. On-device AI assistants went from impossible to real.
John Jeong - Product design and sales. Passionate about simplicity and thoughtful UX. Products should remove friction, not add it.
Yujong Lee - Open-source engineer, contributor to Hono and LiteLLM. Systems thinker who goes deep into fundamentals.
Different backgrounds, shared values: craft, simplicity, building things people actually want to use.
Most AI tools force you into their ecosystem, their models, their rules. Locked into subscriptions, dependent on cloud services, no real control. Hyprnote proves another path exists.
What Makes Us Different:
- Model flexibility - On-device, cloud, or mix and match
- Data ownership - Local-first, privacy-first
- Open source - Transparent, extensible, community-driven
- Future-proof - Built for the shift to local AI
You shouldn't need enterprise budgets to have powerful, flexible AI tools that respect your autonomy.
Hyprnote today is a notetaker. Tomorrow, much more.
- A trusted executive assistant - Notes, inbox, calendar, workflow
- A suite of open-source, AI-native tools - Empowering without lock-in
- Fortress-grade enterprise - Compliance-ready yet simple
Build the AI-native productivity stack that feels as natural as having a world-class assistant by your side.
Not marketing slogans. Convictions we return to when making hard decisions.
Files Endure, Interfaces Change - The filesystem isn't a relic - it's how humans organize information. Location is recall. Structure is memory. Your notes should survive us.
Ownership Over Convenience - A note-taking system is a memory prosthetic. Memory needs to feel owned. Files on your disk, not opaque blobs behind an API.
Compatibility Over Dominance - We don't want to be the only tool you use. Best at meeting notes, speaks the same language as everything else. Plain text makes this possible.
Durability Over Trendiness - Technologies that last decades, not ones popular this quarter. Local-first, Git-backed. Fast for humans, safe for decades.
Transparency Creates Trust - Open source isn't marketing. It's being inspectable. See how it works, trust it. Modify it, own it.
Specialization Within Integration - Don't try to be everything. Best at meeting notes, integrated with your broader system. No walled garden. One thing, done exceptionally well.
Simplicity Creates Coherence - Startups win through clearer philosophy, not feature accumulation. Remove features more often than add them. Minimum complexity for the current task.
Multiple Interfaces, Unified Data - Different thinking modes deserve different environments. Specialized interfaces on unified data. Separation without fragmentation.
Friction-Free Publishing - Writing and publishing merge into one action. Reduced friction removes barriers to documentation.
Substrate Over Abstraction - Filesystem remains foundational as intelligence layers expand. Build on lasting infrastructure. AI evolves, your data remains.
AI Needs Inspectable Artifacts - To help humans think, AI needs artifacts humans can own, inspect, move, and understand. Plain text enables collaboration without dependency.
Legibility Enables Intelligence - AI and humans thrive with simple, transparent systems. The filesystem becomes the cortex - organized, inspectable structures enable understanding and reasoning.
Local-First AI - Hardware advances rapidly. What required cloud now runs on your laptop. On-device AI gives speed, privacy, and independence from subscriptions.
The Human Remains Central - AI reduces cognitive load, not replaces judgment. A tool that works for you, not a system you work inside. Always see what's happening, always override.
Model Agnosticism - No lock-in to our AI models. Bring your own, use local, mix and match. Intelligence layer as flexible as data layer.
We don't differentiate through features. We build a brand.
Open Source Is a Brand - People who love open source often cite transparency, security, or ownership. But at its core, it's preference - like choosing Philz over Starbucks. Open source is already a brand itself, and we embrace that identity fully.
Cloud Isn't the Differentiator - From a user's perspective, most AI notetakers are built on similar infrastructure - the same API providers, the same underlying models. Users don't choose based on whose cloud is better. They choose based on the experience, the values, and the trust they feel.
Branding Is Reputation - It's how people think about us. Not what we say we are, but what they believe we are. We earn that through consistency, transparency, and building something people genuinely want to use.
Values Over Feature Lists - We use terms like control, ownership, no lock-in, no strings attached. These aren't marketing copy - they're the principles that guide every decision. The features follow from the values, not the other way around.
Maximize focus, creativity, and speed. We work from anywhere and embrace that each founder has a different style. Monthly in-person planning, otherwise remote.
John - Nocturnal, moves between coffee shops and different environments. Cardio in morning, weights at night. Organized flexibility.
Yujong - Anti-nocturnal. Wakes early, sleeps early. Calm, ordered environments for deep focus and systems thinking.
Productivity comes not from forcing uniformity, but from creating conditions where each person delivers their best.
Balance independence with sync. Too much disturbance kills focus; too little causes drift.
Tools:
- GitHub Issues and Projects - Single source of truth. Decisions, tasks, prioritization in GitHub Projects. Transparent, no digging through chat.
- Slack - Lightweight coordination. Long updates belong in GitHub. Huddles for real-time discussion.
- Devin - AI software engineer with dedicated Slack channel. Treated like a colleague, transparent workflow.
- GitHub as CMS - IDEs double as blog editors. Write something? Submit a PR.
- Dev tools - Zed, GitButler, Cursor, Claude Code.
Escalation:
- Phone - Truly urgent, can't wait
- Slack mention - Reasonably urgent, can wait hours
- GitHub - Everything else, async
AI-Augmented: Small, mission-aligned team augmented with AI can outperform companies with hundreds of people.
One source of truth, open to everyone. Early on we had too many channels. Now GitHub is primary, everything syncs into it.
GitHub Structure:
- Discussions - Feedback requests, feature conversations
- Issues - Bug reports, tracked work
- Projects - Prioritization, operational board (link)
Focused cycles. Structure with flexibility.
Stable vs Nightly:
- Stable - Polished, user-facing. Each release should make the app "sticky" with genuine improvement.
- Nightly - Internal testing, rapid iteration.
Cadence: Monthly in-person planning. Define focus and outcomes for next stable. Kanban-style board in GitHub Projects gives visibility. Weekly sprints within larger cycles. Clear requirements and deliverables.
Principles:
- One core feature per stable - Anchor each release with one significant improvement
- Habit loop focus - Make Hyprnote indispensable before expanding
- Huddles for discussion - Real-time collaboration over lengthy async threads
- Ship and iterate - Good and shipped beats waiting for perfect
Grit - Building a startup is hard. We need people who push through with resilience and purpose, not blindly, but committed to the mission when progress feels slow.
Intellectual Honesty - Say "I don't know" or "I was wrong." Seek truth over being right. Challenge ideas constructively. Update views with better evidence.
Sharpness - Think quickly, communicate clearly, execute decisively. Not about being smartest - about being effective under pressure and making good decisions fast.
A Vivid Vector - We think of people as a collection of vectors. You don't need to care about everything we care about. But you need a vivid vector that aligns with at least one of our pillars. Artisans with clear direction and genuine passion. Not people with no signal, no taste, no direction.
Hard Skills Matter Too: Soft traits are essential, but we need excellence in what you do. Engineering, design, sales, operations - deep competence. Small team, everyone pulls their weight.
The Example We Set: We hold ourselves to the same standards. Not perfect, but we try to embody grit, honesty, and sharpness every day. Expect future teammates to do the same - and call us out when we fall short.
What You'll Gain: Become a craftsman. The founders have strong taste. You'll be pushed to develop sharp perspectives and refine your craft alongside a small team that cares deeply.
Own your work. Move fast. Communicate openly. We don't micromanage - we trust you to figure things out. In return: autonomy, impact, and the chance to build something meaningful.
How We Work Together: Remote-first. Async via Slack. Occasional calls with Hyprnote on. Everything runs through GitHub - bug reports, feature requests, code review, and project management. Small team, high trust, craft over hype.
People who believe they deserve better than vendor lock-in and subscription treadmills. They want powerful AI tools on their terms.
Developers First: If we had to pick one target audience, it would be developers. We're devs ourselves, and every value we hold was centered around the fact that we love dev tools and products that devs love. The more dev-related taste someone has, the more appealing Hyprnote becomes for them.
Core Traits:
- Autonomy-first - Choose your own AI models, on-device or cloud. Many already have API credits and wonder why they should pay twice for tools like Otter when they can bring their own key.
- Data-conscious - Their notes stay theirs, not training data. They understand the value of local-first architecture.
- Tech-forward - Excited about local AI, want tools that harness it. They appreciate the filesystem as a foundation, not a relic.
- Open source advocates - Value transparency and extensibility. They're part of the open source community not just for practical reasons, but because it aligns with how they think software should be built.
We give people back their focus and autonomy. Be fully present in conversations with complete control over your AI and data.
Core Value - Bring Your Own Key: If you already have API credits from OpenAI, Anthropic, or other providers, why pay twice? Hyprnote lets you use your existing keys instead of forcing you onto another subscription. This builds trust and reinforces everything we stand for.
What You Get:
- Freedom of choice - Any AI model. On-device, cloud, or both. No lock-in to our models.
- True ownership - Your data stays yours. Files on your disk, not opaque blobs behind an API. Local-first means you're never dependent on our servers.
- Time savings - No manual note-taking or post-meeting docs. Focus on the conversation.
- Better recall - Searchable transcripts, nothing gets lost. The filesystem becomes your memory.
- Simplicity - Powerful features that feel effortless. One thing done exceptionally well.
This handbook is a living document. We update it as we learn and grow.