Implementation Guide

Official documentation for developers, archivists, and researchers integrating with the 1990 Web Archive infrastructure.

This guide covers API authentication, crawler configuration, legacy rendering parameters, and data export protocols. All endpoints use secure TLS 1.3 and support both JSON and WARC payloads.

⚡ Platform Notice

Legacy HTTP endpoints (pre-2018) have been deprecated. Use the `v2/` namespace for all new integrations. Migration tools are available in the CLI.

Quick Start

1. Install the CLI

bash

npm install -g @1990archive/cli
archive login --token YOUR_API_KEY

2. Run your first crawl

bash

archive crawl --url "http://geocities.com/area/51/" \
  --depth 2 \
  --engine netscape3.0 \
  --output ./archive-geo-1995.warc

Verify the output using the built-in validator:

bash

archive validate ./archive-geo-1995.warc
[✓] 42 resources captured
[✓] Integrity checksum: a7f9c2d...

Authentication

All API requests require a Bearer token in the `Authorization` header. Tokens are scoped to your organization and expire after 30 days by default.

Scope	Access Level	Rate Limit
`read:archive`	View & download public records	1000 req/min
`write:crawl`	Submit custom crawl jobs	50 jobs/hr
`admin:render`	Custom emulation configs	Unlimited*

* Enterprise tier required for admin scopes.

curl

curl -X GET https://api.1990archive.org/v2/search \
  -H "Authorization: Bearer sk_live_90web_..." \
  -H "Content-Type: application/json"

Crawler API

Configuration Schema

Crawl jobs accept a JSON payload. Key fields include `url`, `depth`, `engine`, and `filters`.

json

{
  "url": "http://www.example.com/1996/",
  "depth": 3,
  "engine": "netscape3.0",
  "filters": {
    "include": [".html", ".gif", ".mid"],
    "exclude": ["javascript:*"]
  },
  "respect_robots": false,
  "legacy_mode": true
}

🔍 Legacy Mode

When `legacy_mode` is true, the crawler bypasses modern TLS handshake requirements and negotiates using SSLv3/TLS 1.0 protocols where necessary for authentic 1990s infrastructure emulation.

Rendering & Emulation

The rendering engine supports deterministic playback of historical user agents. Available profiles:

Netscape Navigator 3.0 (1996) - Table layouts, early CSS, MIDI autoplay
Internet Explorer 3.02 (1996) - ActiveX quirks, proprietary DHTML
Opera 3.5 (1998) - Lightweight rendering, early frames support

bash

archive render --warc ./site.warc \
  --profile ie3.02 \
  --resolution 800x600 \
  --font-fallback "Times New Roman" \
  --screenshot ./output.png

Data Export

Archives can be exported in multiple formats depending on your use case:

Format	Best For	Extension
WARC (v1.1)	Long-term preservation	`.warc.gz`
HTML Bundle	Local browsing	`.zip`
Memento JSON	Time-travel APIs	`.json`
CSV Metadata	Analytics & indexing	`.csv`

Use the `--format` flag to specify output type. Large exports are streamed via signed URLs with 24-hour expiration.

Rate Limits & Best Practices

The API enforces rate limits to preserve infrastructure stability. Always implement exponential backoff.

Monitor the X-RateLimit-Remaining header
Use Accept-Encoding: gzip to reduce bandwidth
Cache WARC responses locally when possible
Avoid crawling dynamic JS-heavy pages in legacy mode

javascript

async fetchWithBackoff(url) {
  let retries = 3;
  while (retries > 0) {
    const res = await fetch(url);
    if (res.status !== 429) return res;
    await sleep(1000 * (3 - retries));
    retries--;
  }
}

Troubleshooting

Error	Cause	Resolution
`E_TLS_LEGACY`	Target server dropped SSLv3	Disable `legacy_mode` or use proxy
`W_RENDER_TIMEOUT`	Heavy Flash/MIDI load	Reduce resolution or disable media
`E_404_MEMENTO`	Timestamp out of archive range	Verify `memento-datetime` header
`W_MIME_MISMATCH`	Server sends incorrect content-type	Archive will auto-detect & warn

Still stuck? Open an issue on GitHub or contact dev-support@1990archive.org with your request ID.