Product
A picture is worth a thousand words. Especially after a meeting.
Why people remember diagrams better than text notes, how Mermaid lets you write diagrams as code, and how AI (and Neural Summary) draws them for you.
May 16, 2026 · 7 min read
Engineering
Five languages from day one: internationalization in a full-stack monorepo
Why we built internationalization into the first commit, how we handle three layers of language (UI strings, AI-generated content, on-demand translation), and what we would do differently.
May 14, 2026 · 8 min read
Engineering
Dark mode, skeletons, and micro-animations: the details that make software feel alive
Why we invested a week in polish — preventing the dark mode white flash, replacing every spinner with layout-matched skeletons, and adding GPU-composited micro-animations. What each decision cost and what it changed.
May 12, 2026 · 7 min read
Thought Leadership
The meeting notes market is solving the wrong problem
The meeting AI space is crowded with tools competing on transcription quality. But the actual pain point is not 'I don't know what was said.' It is 'I still have to write the deliverable.' The market is optimizing the wrong phase.
May 9, 2026 · 7 min read
Product
How to take notes in Google Meet (and ship follow-ups in minutes)
How to take notes in Google Meet without losing the action. A practical 3-step guide: safeguard the audio, generate clean structured notes, and turn conversations into work artifacts. Works on every Google Meet plan, including free Gmail.
May 8, 2026 · 13 min read
Engineering
Real-time progress for long-running AI tasks
Why we moved transcription processing off the request cycle, added WebSocket with polling fallback, and how those decisions shaped our architecture.
May 7, 2026 · 7 min read
Product
The mobile companion: recording on the go
How we built a cross-platform mobile app with Expo in five phases. Authentication sharing, upload queues, quota-aware retry logic, and the engineering behind graceful plan limits.
May 5, 2026 · 7 min read
Product
30 lens templates and counting: how we design AI-powered document generators
The five-category value ladder, the anatomy of a lens template, and how we evolved from generic summaries to consulting-grade output across 30 document types.
May 2, 2026 · 8 min read
Engineering
Eight security vulnerabilities we found in our own code
We audited our own codebase before launching V2 and found eight vulnerabilities, including a critical command injection. What we found, why each one existed, and the architectural patterns we adopted to prevent them.
April 30, 2026 · 7 min read
Engineering
Building streaming AI chat with Server-Sent Events
Why we chose SSE over WebSocket for AI chat streaming, how citation buffering eliminated visual artifacts, and the architectural decisions behind multi-turn context and stop generation.
April 28, 2026 · 6 min read
Product
Folder intelligence: asking questions across 50 conversations at once
How we built cross-conversation AI chat with semantic search, streaming responses, and citations that reference specific timestamps and speakers across different conversations.
April 25, 2026 · 6 min read
Thought Leadership
Three hundred million meetings happen every day
The old benchmark of eleven million meetings per day is more than a decade out of date. The modern number is closer to three hundred million. And almost none of them produce a finished deliverable.
April 25, 2026 · 8 min read
Engineering
Prompt engineering for multilingual structured output
Why we moved from markdown to structured JSON output, and the prompt patterns we developed over 30 templates to get consistent quality in five languages.
April 21, 2026 · 9 min read
Engineering
Processing 5GB audio files: building a resilient media pipeline
Why we built an audio splitting pipeline, what broke when real users started uploading browser recordings, and the fallback layers that made the system resilient.
April 18, 2026 · 6 min read
Product
I built an AI meeting notes platform in 6 months. Here's the full timeline.
From first commit to streaming AI chat, 671 commits later. The decisions that paid off, the ones I'd make differently, and what I learned building a full-stack AI product as a solo founder.
April 10, 2026 · 9 min read
Thought Leadership
Most meeting AI optimizes the wrong 60 minutes
The real productivity bottleneck is not the meeting. It is the 90 minutes afterward, translating decisions into deliverables. Most meeting AI focuses on the wrong phase.
March 30, 2026 · 7 min read
Thought Leadership
Meeting summaries are not deliverables
Most meeting tools stop at telling you what was said. The real value is generating the deliverable you would have written manually: the spec, the brief, the email, the backlog.
March 20, 2026 · 6 min read
Consulting
The two hours after every client meeting are killing your margins
Independent consultants spend 25% of their week on post-meeting write-ups. That is where margins shrink. There is a faster way.
March 10, 2026 · 5 min read
Product
Why knowledge workers spend more time writing than thinking
The thinking was done in minutes. The document took hours. That translation step is where momentum dies, and it does not have to be manual.
March 4, 2026 · 5 min read