Granola is an AI-powered notepad designed for the constraint of back-to-back meetings. The core problem it addresses: users cannot both stay present in a call and produce usable notes afterward. The product approach is direct - transcribe audio passively without requiring a bot in the meeting, allow manual note-taking to continue in parallel, then apply enhancement (summarization, rewriting, structural improvement) post-meeting to convert raw input into actionable output.
The technical stack spans audio transcription, natural language processing for note enhancement, searchable document indexing, and a chat interface that maintains context across a user's meeting history. The product positions itself as Apple Notes with transcription - a familiar editing surface rather than a specialized meeting tool. iPhone support is included.
The value proposition is throughput: teams move faster when meeting discussions are automatically converted into polished, indexed outputs. The latency trade-off is explicit - enhancement happens after the call, not during, which preserves user presence but requires post-processing. The operational model is simple: no meeting integrations or bot infrastructure to manage; audio flows from device to the product directly.
Granola is headquartered in London and remains privately held.