BB Recorder

Name: BB Recorder — Download page. Local Recording and Private Transcription
Availability: InStock
Rating: 5 (1 reviews)
Author: BuildBetter

BB Recorder — Free Download. Local Recording and Private Transcription

BB Recorder is a meeting and call recording application that operates entirely on the users device. Originally developed for macOS, the program captures system and microphone audio, transcribes conversations using on-device artificial intelligence, and generates automatic summaries without sending data to external servers. The application works with Apple Intelligence, local Whisper models, or API keys provided by the user.

★★★★★

5.0(1 ratings)

Download BB Recorder (Official links)

File size: 22 MB

The latest version of BB Recorder is: 1.0.0

Operating system: MacOS, iOS

Languages: Spanish, English

Developer: BuildBetter

Price: $0.00 USD

NOTCH-style menu bar recording. The program integrates a compact panel in the top menu bar that expands visually like the iPhone's notch. This mode occupies minimal space, displays real-time audio meters, and allows recording control with a single click. The visual indicator changes color during capture.
Floating overlay. A semi-transparent window that remains on top of other applications. The overlay shows duration, a waveform, and pause buttons. It can be repositioned anywhere on the screen and resized. Ideal for recording sessions while working with other tools.
Full window mode. A dedicated interface with live transcription display, a notes editor, and a summaries panel. This view allows managing multiple simultaneous recordings and accessing the complete history. Includes a navigable timeline.
Speaker diarization. Automatic identification of participants through voice pattern analysis. The system assigns distinct colors to each speaker in the transcription. The algorithm differentiates voices without requiring prior training or manual identification. Accuracy increases with longer recordings.
Local transcription with Whisper. Full implementation of OpenAI models running on the local machine. No audio fragments are transmitted to the cloud. Supports tiny, base, small, medium, and large sizes. Processing speed depends on available hardware. Audio files in WAV, M4A, MP3, and AAC formats.
Automatic summarization with Apple Intelligence. On macOS Sequoia and later, BB Recorder uses the unified neural engine to generate conversation extracts. The summary identifies action items, agreed-upon dates, and decisions made. Works offline.
External API key integration. Allows configuration of cloud services using your own credentials. Compatible with Deepgram for low-latency transcription, OpenAI GPT-4 for advanced summaries, and Anthropic Claude for contextual analysis. The user retains full control over usage and billing.
Live assistant during calls. A voice or text query function while recording is active. The user can ask questions like what points have been covered so far. The model processes the transcribed segment and responds by synthesizing the available information. The answer appears in an overlay window.
Intelligent call detection. The program monitors system audio activity. Upon detecting the start of a call on FaceTime, Zoom, Teams, Slack, or Google Meet, it automatically activates recording. Activation thresholds are configured per application or by sound pressure level.
Notes editor with Markdown syntax. A side panel with full support for headers, lists, code blocks, links, and images. Split view between editor and HTML preview. Notes are automatically linked to the corresponding transcription file. Titles are generated from the first words of the recorded content.
Multi-channel export. Transcription is saved in TXT, SRT, VTT, and PDF formats. The original audio file can be exported in lossless compression. The export package includes metadata: date, duration, detected participants, and the transcription model used.
Upload to BuildBetter for teamwork. Optional connection to the BuildBetter platform via user authentication. Files are transferred encrypted for collaborative analysis. This feature is not active by default and requires explicit activation.

The first version of BB Recorder was released in February 2025. The author conceived the tool after verifying that existing solutions uploaded recordings of confidential meetings to servers in the United States without end-to-end encryption. The original program was written in Swift 5.5 using the AVFoundation framework for audio capture and Core ML for the first experiments with on-device recognition. In 2023, the transcription engine was partially rewritten to integrate Whisper.cpp bindings, implemented in C++ and Objective-C++. The current version retains the Swift core in the presentation layer and asynchronous processing.