Yes. Condense is designed from the ground up to be stateless and operate completely in memory using Buffers and Streams. It avoids temporary disk writes by default, eliminating I/O bottlenecks and race conditions in concurrent or serverless environments.
Yes. Condense v1.0.0 includes a built-in WorkerPool that leverages Node.js worker_threads to distribute CPU-intensive operations (image encoding, perceptual SSIM calculations, ZIP processing) across multiple worker threads.
Perceptual optimization uses the Structural Similarity Index (SSIM) and Peak Signal-to-Noise Ratio (PSNR) to measure perceived visual degradation. Rather than picking an arbitrary compression quality number, Condense performs a binary search to find the absolute minimum file size that satisfies your exact visual quality requirement (e.g. SSIM
No. The default design is strictly in-memory processing. Temporary disk writes are avoided unless explicitly required by specific operations (such as MP4 faststart).
quality: Best fidelity with visually lossless, conservative compression. Ideal for photography or brand assets.balanced: A practical balance of file size and speed. Recommended default for standard web delivery.extreme: Maximum compression for bandwidth-critical contexts, converting assets to modern formats and stripping non-critical metadata.