Product· 8 min read
How we built MCP context injection for meeting intelligence
A deep dive into how Reverbic turns your last three weeks of meeting history into structured context that AI coding agents can actually use.
Blog
Writing from the team.
Research· 5 min read
The 37-minute cost: what bad meeting notes actually cost you
We surveyed 800 knowledge workers and found that the average engineer spends 37 minutes per week trying to reconstruct context from meetings they attended.
Engineering· 11 min read
Inside our Whisper post-processing stack
Whisper is great. Our custom post-processing layer — speaker normalisation, domain vocabulary boosting, and confidence-weighted re-scoring — is what sharpens transcription accuracy.
Product· 6 min read
Decision Tracker: why your team keeps re-litigating old decisions
The average team re-makes the same architectural or strategic decision 2.3 times. Here's how we built a searchable log to stop that.