Open Research
Open Research
We publish our evaluation methodology, prompt designs, and model comparisons so you can see exactly how BrainTube works under the hood.
Transcription quality benchmark
How we compare Whisper variants, AssemblyAI, and Deepgram across accents, audio quality, and domain vocabulary.
Writeup coming soon
Summarization eval framework
Faithfulness, coverage, and compression metrics across Claude, GPT, and Gemini for long-form video.
Writeup coming soon
Semantic search evals
Recall@k benchmarks across embedding models for transcript chunks.
Writeup coming soon