An indie hacker just made Granola free and open source

Signals Inbox·July 27, 2026·AI Productivity

An indie developer has released Quill, an open-source meeting notetaker inspired by Granola. It captures meeting audio, creates a live transcript, and turns rough notes into a polished write-up. The tension is simple: Granola charges $14 per month for unlimited history and advanced features, while a developer has rebuilt much of the core workflow as an MIT-licensed app. Quill is still pre-alpha and requires paid AI API keys, but it shows how quickly a popular AI product can become an open-source feature set.

Q1What actually launched?

Quill is an open-source meeting notetaker published on GitHub. You open it before a meeting, press record, and take rough notes while it captures your microphone and computer audio. When the meeting ends, it combines your notes and transcript into a cleaner write-up using templates. The code uses an MIT license, which means other developers can inspect, modify, and redistribute it.

Q2Is it really a free Granola?

It copies a large part of the basic workflow, but it is not yet a complete replacement. Like Granola, Quill records without adding a bot to the call, produces a transcript, and turns rough notes into something readable. But Granola is a mature product available across desktop and mobile, with team features, integrations, shared folders, AI chat, and support. Quill is currently a pre-alpha Mac app that users must install and configure themselves.

Q3How much money could it save?

Granola has a free plan, but it only shows limited meeting history. Its Business plan costs $14 per user each month and adds unlimited notes, advanced AI models, integrations, API access, and team management. Quill does not charge a subscription or license fee. However, users currently need their own OpenAI and Anthropic API keys, so the AI processing is not literally free. You replace a fixed subscription with usage-based API costs and more setup work.

Q4Is Quill completely private?

Not in its current form. Quill says it does not keep the original audio after processing, and its notes are stored locally in a SQLite database. But it uploads audio in 20-second chunks to OpenAI Whisper for transcription and sends notes to cloud AI models such as Claude for enhancement. Granola also uses outside transcription and AI providers and stores transcripts and notes in its cloud. Quill gives users more control over the code and storage, but the current version is not fully offline.

Q5Why does this matter now?

Because AI application features are getting easier to reproduce. The difficult building blocks already exist: system-audio capture, Whisper transcription, Claude or GPT for summaries, and open-source desktop frameworks. Quill reached a working pre-alpha repository with only a small development history. That does not make it better than Granola, but it shows that transcription and note generation alone may not remain a strong moat.

Q6So is Granola in trouble?

Not because of one early GitHub project. Granola still has a major lead in product polish, reliability, integrations, mobile access, distribution, and team workflows. The bigger warning is about where its long-term advantage must come from. If basic recording and summarization become cheap or free, Granola has to win through better memory, search, collaboration, security, and connections to the rest of a company’s software. Quill is not the winner yet. It is proof that the basic feature is becoming easier to copy.