How does API payload size affect cloud bandwidth and egress costs?

Every byte returned by your API endpoints traverses cloud network providers. At scale, transmitting uncompressed or bloated JSON with redundant nulls, verbose timestamps, or repeated nested objects multiplies data transfer costs ($0.05 - $0.12 per GB) and increases client network latency.

What is the difference between Gzip, Brotli, Zstandard, MessagePack, and CBOR for API payloads?

Gzip is universally supported by web clients. Brotli provides 15-25% better compression for text/JSON. Zstandard (Zstd) offers near-Brotli ratios at much higher decompression speeds. MessagePack and CBOR are binary serialization formats that reduce payload size before compression by eliminating JSON string parsing overhead.

Is my API payload data uploaded or logged by BandwidthGuard?

No. BandwidthGuard runs 100% client-side inside your browser. No JSON payloads, headers, or API tokens are uploaded, tracked, or sent to any server.

How can I reduce my API response payload size?

You can reduce API response size by: 1) enabling Gzip or Brotli compression, 2) stripping null fields, 3) using Unix timestamps instead of ISO 8601 strings, 4) paginating lists to 20-50 items, 5) converting uniform object arrays to columnar arrays, 6) truncating floating point precision to 4 decimals, and 7) utilizing sparse fieldsets.

BandwidthGuard — Stop paying for bloated API payloads

What is this body?

Request = what a client sends to the backend. Response = what the API returns. Advice and copy-optimize follow this choice.

Response body

Valid JSON — returned by the API

{
  "page": 1,
  "total": 128,
  "results": [
    {
      "id": "usr_1000",
      "full_name": "Sample User 1",
      "email_address": "user1@example.com",
      "created_at": "2026-01-01T09:30:00.000Z",
      "updated_at": "2026-06-01T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1001",
      "full_name": "Sample User 2",
      "email_address": "user2@example.com",
      "created_at": "2026-01-02T09:30:00.000Z",
      "updated_at": "2026-06-02T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1002",
      "full_name": "Sample User 3",
      "email_address": "user3@example.com",
      "created_at": "2026-01-03T09:30:00.000Z",
      "updated_at": "2026-06-03T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1003",
      "full_name": "Sample User 4",
      "email_address": "user4@example.com",
      "created_at": "2026-01-04T09:30:00.000Z",
      "updated_at": "2026-06-04T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1004",
      "full_name": "Sample User 5",
      "email_address": "user5@example.com",
      "created_at": "2026-01-05T09:30:00.000Z",
      "updated_at": "2026-06-05T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1005",
      "full_name": "Sample User 6",
      "email_address": "user6@example.com",
      "created_at": "2026-01-06T09:30:00.000Z",
      "updated_at": "2026-06-06T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1006",
      "full_name": "Sample User 7",
      "email_address": "user7@example.com",
      "created_at": "2026-01-07T09:30:00.000Z",
      "updated_at": "2026-06-07T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    },
    {
      "id": "usr_1007",
      "full_name": "Sample User 8",
      "email_address": "user8@example.com",
      "created_at": "2026-01-08T09:30:00.000Z",
      "updated_at": "2026-06-08T09:30:00.000Z",
      "avatar_url": null,
      "last_login_ip": null,
      "preferences": {
        "theme": "dark",
        "locale": "en-US",
        "newsletter": false
      }
    }
  ]
}

Samples:

Pretty printing costs +904 B vs minified (2.13 KB).

Size analysis

Measured in your browser, nothing uploaded

Raw

3.01 KB

3,082 bytes

Gzip

0 B

0 bytes

Brotli (est.)

0 B

0 bytes

Kilobytes3.01 KB

Megabytes0.0029 MB

Gzip ratio1% smaller

Brotli ratio1% smaller

52score

Trimmable

Bloat score weighs whitespace, null fields, base64 blobs, verbose timestamps and repeated keys against the total payload size.

Strings1.91 KBNumbers4 BObjects108 BArrays9 BWhitespace904 BOther104 B

Serialization & Compression Matrix

Live benchmark across text, binary, and stream compression codecs

Raw JSON

Text

Uncompressed original payload

3.01 KB

3,082 bytes on wire

Minified JSON

Text

Whitespace stripped

2.13 KB−29%

2,178 bytes on wire

MessagePack

Binary

Standard compact binary JSON equivalent

0 B

0 bytes on wire

CBOR

Binary

Concise Binary Object Representation (RFC 8949)

0 B

0 bytes on wire

Gzip (Level 6)

Compressed

Ubiquitous HTTP/1.1 & HTTP/2 compression

0 B

0 bytes on wire

Brotli (est.)

Compressed

Modern web standard, higher ratio for text

0 B

0 bytes on wire

Zstandard (est.)

Compressed

Next-gen HTTP/3 compression — estimated from Brotli ratio

0 B

0 bytes on wire

Field-level size breakdown

Top JSON fields and nested paths contributing the most bytes to your payload

Field / Path	Count	Size	% of Total
results	×1	2.10 KB	70%
results[*].preferences	×8	528 B	17%
results[*].created_at	×8	312 B	10%
results[*].updated_at	×8	312 B	10%
results[*].email_address	×8	280 B	9%
results[*].full_name	×8	216 B	7%
results[*].last_login_ip	×8	160 B	5%
results[*].preferences.newsletter	×8	144 B	5%
results[*].avatar_url	×8	136 B	4%
results[*].preferences.locale	×8	128 B	4%

Traffic simulator

Model your monthly load

Monthly active users50,000

10010M

Requests per user per day20

1100

Compression on the wire

What-if scenarios

Monthly bandwidth

86.11 GB

3082 bytes per response

Monthly requests

30.00M

20/user/day × 30 days

Estimated peak RPS

uniform distribution, 2× peak factor

Mobile & Network Latency Simulator

Estimated download times for 3.01 KB on the wire, including TCP slow-start round trips

Slow 3G

Emerging markets / weak mobile signal (400 Kbps, 400ms RTT)

462 ms1 RTT

Transfer: 62msRTT: 400ms

Fast 3G

Typical 3G connection (1.6 Mbps, 150ms RTT)

165 ms1 RTT

Transfer: 15msRTT: 150ms

4G / LTE

Standard mobile LTE (12 Mbps, 50ms RTT)

52 ms1 RTT

Transfer: 2msRTT: 50ms

5G / Fiber

High-speed broadband or 5G (100 Mbps, 15ms RTT)

15 ms1 RTT

Transfer: 0msRTT: 15ms

Satellite / Starlink

Satellite link with propagation delay (25 Mbps, 220ms RTT)

221 ms1 RTT

Transfer: 1msRTT: 220ms

Monthly cost comparison

Estimates · 86.11 GB egress · best value is per architecture group

Region:

Provider	Plan	Included	Egress cost	Request cost	Total monthly	Per 1M requests
Edge & CDN
Cloudflare CDNBest Value	CDN (Free / Pro / Biz zones)	Unmetered CDN egress	$0.00 $0.00 / GB	$0.00 $0.00	$0.00	$0.00
Azure CDN	Microsoft Standard	None	$6.97 $0.081 / GB	$0.00 included	$6.97	$0.23
GCP Cloud CDN	Cache egress	None	$6.89 $0.080 / GB	$22.50 $0.0075 / 10k	$29.39	$0.98
AWS CloudFront	Pay-as-you-go	1 TB / month (Always Free)	$0.00 $0.085 / GB	$30.00 $0.0100 / 10k	$30.00	$1.00
Fastly	Pay-as-you-go	$50 free trial credit	$10.33 $0.120 / GB	$22.50 $0.0075 / 10k	$32.83	$1.09
Frontend & Serverless Cloud
VercelBest Value	Pro	1 TB Fast Data Transfer	$0.00 $0.150 / GB	$0.00 not modeled	$0.00	$0.00
Netlify	Pro ($19/seat)	1 TB Bandwidth included	$0.00 $0.200 / GB	$0.00 included	$0.00	$0.00
Azure Static Web Apps	Standard	100 GB included / month	$0.00 $0.087 / GB	$0.00 included	$0.00	$0.00
Supabase	Pro ($25/mo)	250 GB unified egress	$0.00 $0.090 / GB	$0.00 included	$0.00	$0.00
Railway	Hobby / Pro	$5 free credit (~50 GB)	$3.61 $0.100 / GB	$0.00 included	$3.61	$0.12
Render	Pro workspace	25 GB included	$9.17 $0.150 / GB	$0.00 included	$9.17	$0.31
Backend Origin & VPS
AWS EC2 / BackendBest Value	Internet Egress	100 GB / month (Always Free)	$0.00 $0.090 / GB	$0.00 included	$0.00	$0.00
Azure VMs / Backend	Internet Egress	100 GB / month (Always Free)	$0.00 $0.087 / GB	$0.00 included	$0.00	$0.00
GCP Compute / Cloud Run	Internet Egress	100 GB / month (Always Free)	$0.00 $0.120 / GB	$0.00 included	$0.00	$0.00
DigitalOcean	Droplets ($6/mo)	1,000 GB included	$0.00 $0.010 / GB ($10/TB)	$0.00 included	$0.00	$0.00
Hetzner Cloud	Cloud CX	20 TB included (EU) / 1 TB (US)	$0.00 $0.001 / GB (€1/TB)	$0.00 included	$0.00	$0.00
Fly.io	Pay as you go	None (new orgs)	$1.72 $0.020 / GB	$0.00 included	$1.72	$0.06

Estimates, not invoices. List prices verified 2026-08-23 for US & Europe. Excludes compute, storage, seats, and enterprise discounts. CDN, serverless, and origin hosts are grouped separately as their architectures and pricing structures differ.

Cloudflare CDN · 2026-08-23
AWS CloudFront · 2026-08-23
Fastly · 2026-08-23
GCP Cloud CDN · 2026-08-23
Azure CDN · 2026-08-23
Vercel · 2026-08-23
Netlify · 2026-08-23
Azure Static Web Apps · 2026-08-23
Supabase · 2026-08-23
Railway · 2026-08-23
Render · 2026-08-23
AWS EC2 / Backend · 2026-08-23
Azure VMs / Backend · 2026-08-23
GCP Compute / Cloud Run · 2026-08-23
DigitalOcean · 2026-08-23
Hetzner Cloud · 2026-08-23
Fly.io · 2026-08-23

Optimization recommendations

This is an API response — do not change the contract unless clients can accept it

Minify is always included (same JSON, less whitespace). Contract changes stay off until you tick them. Gzip/Brotli, binary formats, and key renames cannot be applied in the copied text — change the server or serializer.

Minify JSON

−904 B · 29.3%

Strip indentation and newlines. Same data — this is always applied when you copy.

Always applied

Convert repeated objects to columnar JSON

−714 B · 23.2%

results has 8 uniform items. Converting to {columns, rows} eliminates key repetition.

Columnar encoding replaces object keys on every row with a single header row and rows array. Client deserializers must support this.

Yes — apply this when copying (I confirmed clients allow it)

Don't repeat the same nested object on every row

−528 B · 17.1%

preferences ×8 is identical on each item. Lift it to a default or omit it on the list.

If every row has the same nested object, send it once or omit it on the list. Confirm the client can live without it per row.

Yes — apply this when copying (I confirmed clients allow it)

Keep PII off the list payload

−440 B · 14.3%

Looks like personal data: email_address, last_login_ip. Prefer a detail endpoint instead of repeating them on every row.

List screens rarely need email, IP, or phone. Only drop them if no client reads those keys.

Yes — apply this when copying (I confirmed clients allow it)

Remove null fields

−312 B · 10.1%

16 null values can be omitted entirely.

Clients may still read this key. Only drop it if the UI does not need to tell null apart from a missing field.

Yes — apply this when copying (I confirmed clients allow it)

This list is already paged — slim each row

−308 B · 10.0%

Showing 8 of 128. Don't add more per-row fields; use a detail route for the rest.

Do this in code / on the server

Shorten repeated key names

−272 B · 8.8%

email_address → email, last_login_ip → ip, preferences → prefer. Rename in the serializer — not applied automatically.

Do this in code / on the server

Shorten repeated keys

−247 B · 8.0%

The same object shape repeats many times; key names dominate the payload. Rename in code — not applied automatically.

Do this in code / on the server

Use Unix timestamps instead of ISO 8601

−166 B · 5.4%

16 ISO date strings shrink by roughly 40%.

Only convert if clients already parse numbers. Date pickers and OpenAPI schemas often require ISO strings.

Yes — apply this when copying (I confirmed clients allow it)

Omit false / 0 / empty-string defaults

−144 B · 4.7%

8 default values can be left off if missing means the same thing.

Clients often treat a missing key as false/0/"". Only drop defaults if the UI already does that.

Yes — apply this when copying (I confirmed clients allow it)

Approved rewrite (minified · 0/6 contract changes ticked): 2.13 KB vs 3.01 KB original.

API Optimization & Bandwidth FAQ

Engineering reference and frequently asked questions for API performance