File Size Cheat Sheet for Developers

Why Size Estimation Matters

Estimating file sizes and storage requirements is a daily task for developers. Whether you are choosing between image formats, planning database capacity, setting upload limits, estimating CDN costs, or debugging why a Docker image is 2 GB, you need a mental model of how big things are. This cheat sheet provides the reference points to build that intuition.

All sizes below are approximate and represent typical values. Actual sizes vary based on content, compression, encoding, and quality settings. When precision matters, use our Byte Calculator to convert between units.

Text and Code

Plain text is surprisingly compact. An ASCII character is exactly 1 byte, and a UTF-8 character is 1 to 4 bytes (most English text is 1 byte per character, most CJK characters are 3 bytes).

Rule of thumb: For English text, 1 KB is roughly 1,000 characters or about 150-200 words. A typical API response is under 100 KB; if your endpoint routinely returns more than 1 MB of JSON, consider pagination.

Images

Image file size depends on dimensions, color depth, compression, and format. Uncompressed, a 1920x1080 image at 24-bit color is exactly 6,220,800 bytes (about 6 MB). Compression reduces this dramatically:

Rule of thumb: Serve WebP or AVIF on the web. Target under 200 KB for hero images and under 50 KB for thumbnails. Use responsive images with srcset to avoid sending 4K images to mobile screens.

Audio and Video

Media files are where storage needs escalate quickly. A single minute of uncompressed CD-quality audio (44.1 kHz, 16-bit, stereo) is about 10.5 MB. Video adds orders of magnitude more data.

Rule of thumb: A 1-hour 1080p video is about 4-8 GB compressed. A podcast episode (mono, 64 kbps) is roughly 30 MB per hour. Use the Transfer Time tab to estimate how long media files take to upload or download.

Databases and Data Engineering

Database size estimation is critical for capacity planning, backup strategies, and cost forecasting. The sizes below represent typical values; your specific schema, indexes, and data distribution will affect actual sizes.

• 1 million rows of a typical transactional table: 200 MB to 1 GB
• 1 billion rows: 200 GB to 1 TB (roughly 1,000x the million-row size)
• Index overhead: Typically adds 20-50% to base table size
• WAL/binlog: Can equal or exceed the data size during heavy write periods
• Backups: Compressed backups are typically 3-10x smaller than on-disk size
• A UUID column: 16 bytes per row (128 bits), or 36 bytes as text
• A Unix timestamp (int32): 4 bytes per row. See our Epoch Converter for timestamp formats.

Rule of thumb: For back-of-envelope calculations, estimate 500 bytes per row for a typical OLTP table with 10-15 columns. Add 30% for indexes. Double it for replication overhead.

Developer Artifacts

Understanding the size of common developer artifacts helps with CI/CD optimization, Docker image tuning, and deployment planning:

• A typical npm project (with node_modules): 200-500 MB
• A Go binary (statically linked): 5-20 MB
• A Rust binary (release, stripped): 2-10 MB
• A Java/Spring Boot uber-JAR: 50-150 MB
• A minimal Docker image (Alpine): 5-8 MB
• A Node.js Docker image (Debian-based): 150-300 MB
• A typical .git directory: 10-100 MB (depends on history length)
• Hash output sizes: MD5 = 16 B, SHA-1 = 20 B, SHA-256 = 32 B. See our Hash Generator.

Rule of thumb: If your Docker image exceeds 500 MB, use multi-stage builds. If your .git directory exceeds 1 GB, consider shallow clones in CI.

Cloud Storage Cost Estimates

Cloud storage pricing varies by tier, but approximate rates for 2025-2026 are:

• Object storage (S3, GCS, Azure Blob): ~$0.023/GB/month for standard tier
• Block storage (EBS, Persistent Disk): ~$0.08/GB/month for SSD
• Database storage (RDS, Cloud SQL): ~$0.115/GB/month for provisioned SSD
• CDN bandwidth (egress): ~$0.085/GB for the first 10 TB

At these rates, 1 TB of S3 Standard storage costs about $23/month, while the same data on RDS costs about $115/month. Use the Storage Planner to estimate costs for your specific workload.

Quick Conversion Reference

1 Byte  = 8 bits
1 KB    = 1,000 bytes (SI)  |  1 KiB = 1,024 bytes (IEC)
1 MB    = 1,000 KB          |  1 MiB = 1,024 KiB
1 GB    = 1,000 MB          |  1 GiB = 1,024 MiB
1 TB    = 1,000 GB          |  1 TiB = 1,024 GiB
1 PB    = 1,000 TB          |  1 PiB = 1,024 TiB

Quick math:
  2^10 = 1,024        ≈ 1 thousand
  2^20 = 1,048,576    ≈ 1 million
  2^30 = 1,073,741,824 ≈ 1 billion
  2^40 ≈ 1.1 trillion

For a deeper dive into the SI vs IEC distinction, see our KB vs KiB article.