Savvy
Prepare meeting briefs from local materials and provide real-time source-grounded guidance
About Savvy
Savvy is a local-first meeting assistant for macOS: it reads your own source documents, prepares meeting briefs, and provides source-grounded guidance during live conversations. Preparation can be limited to a specific client or kept general. Documents and history stay on the Mac; real-time transcription requires a connection to Deepgram or AssemblyAI, and suggestions are recommended to be generated through an installed OpenAI Codex CLI or Anthropic Claude Code. It suits macOS users who need to prepare materials before meetings and receive immediate prompts during conversations.
Best for
Best use cases
Prepare a brief before a client meeting
Get source-grounded guidance during a live conversation
Limit preparation to a specific client
Prepare for a general meeting
Features
Source document briefs
Prepare meeting briefs from your own source documents.
Source-grounded guidance
Provide source-grounded guidance in live conversations.
Client-scoped preparation
Limit preparation to a specific client or keep it general.
External transcription and CLI guidance
Connect transcription to Deepgram or AssemblyAI; generation is recommended through an installed Codex CLI or Claude Code.
Pros
- Documents and history are stored on the Mac
- Preparation scope can be set for a specific client or a general use case
Cons
- Real-time transcription depends on Deepgram or AssemblyAI, with meeting audio streamed to the relevant service; there is currently no offline mode
- Requires macOS 13 or later; official builds are Apple Silicon only, while Intel can be built but is unsupported