📚 Crypto Whitepapers Archive

A searchable archive of 880+ cryptocurrency & blockchain whitepapers — every PDF mined for founders, wallet addresses, social links, and publication dates.

🔎 Search 880+ whitepapers live →

Browse the index  ·  Stats  ·  JSON  ·  CSV  ·  Schema

---

What is this?

A snapshot of the crypto whitepaper landscape — 882 PDFs, mostly from the 2016–2018 ICO boom, the single largest wave of whitepaper publishing in history.

Most whitepaper "archives" are just a folder of PDFs. This one is mined: a small, dependency-free toolchain reads every PDF and extracts the metadata you actually want to query —

• 🧑‍💼 Founders & team — role-confirmed names (CEO, CTO, founder, advisor…)
• 💰 Wallet & contract addresses — ETH/EVM and Bitcoin addresses found in the text
• 🔗 Social & community links — Twitter/X, Telegram, Discord, GitHub, Reddit, Medium, LinkedIn…
• 📅 Publication years — inferred from each document

…all dumped into data/whitepapers.json and data/whitepapers.csv so you can grep, query, and analyze the whole ICO era in seconds.

🔬 Why it's useful: researchers studying ICO-era fraud, journalists tracing teams across projects, on-chain analysts mapping addresses to projects, and anyone curious how 882 projects pitched themselves at the peak of the hype cycle.

---

📊 The archive at a glance

Whitepapers ............   882
With extractable text ..   660
Scanned (image-only) ...   222
Dated .................   641
With team names ........   293
With social links ......   230
With wallet addresses ..    39  (66 ETH/EVM, 9 BTC)

Publication years — the ICO boom in one chart:

  2013  █ 5
  2014  █ 7
  2015  █ 11
  2016  ███████ 54
  2017  ████████████████████████████████████████████████ 382
  2018  ███████████████████ 153
  2019   2
  2020  █ 5
  2021  █ 5
  2022   1

🚀 62% of these whitepapers were published in 2017 alone — the peak of ICO mania.

Social platforms linked across the archive:

Platform	Papers	Platform	Papers
GitHub	110	Medium	50
Twitter / X	73	Reddit	18
Facebook	72	Discord	3
Telegram	63	LinkedIn	52

---

🔎 Notable findings (from the extracted data)

Mining the archive surfaces things a folder of PDFs never could:

• Recycled boilerplate. Augmentors and Billionaire Token list the same two Ethereum contract addresses (0x49aec0…17ed7, 0x23EA01…3ff1) — a fingerprint of a shared team or copy-pasted template.
• Example addresses pasted as real ones. Bitcoin's genesis address (1A1zP1…DivfNa, Satoshi's) and a well-known Ethereum whale wallet (0xb794f5…579268) appear across multiple unrelated whitepapers — a reminder to always verify before sending anything anywhere.
• Repeat advisors. A handful of names recur across many projects — e.g. Michael Terpin (4 papers), David Drake (3), Vitalik Buterin (3, mostly as a citation) — the small world of ICO-era advisory.

Run the queries below and you'll find more. PRs documenting interesting findings are welcome.

---

🚀 Quick start

git clone https://github.com/ImMike/Crypto_Whitepapers.git
cd Crypto_Whitepapers

🔎 Search (no install): the live search site filters all 882 papers by name, team, wallet address, social handle or year as you type.

Full-text search, offline (tools/search.py — searches inside the papers):

python tools/search.py "proof of stake"     # full text of every whitepaper
python tools/search.py --wallet 0x49aec0     # which projects list this address
python tools/search.py --team Terpin         # find a person across the archive
python tools/search.py --social telegram --year 2017

Full-text search on GitHub: because each paper's text is committed under text/, GitHub code search searches the whitepaper contents natively.

Browse: open data/INDEX.md for a clickable A–Z table of all 882 papers with their metadata.

Query the JSON (needs jq):

# Every paper that exposes an ETH/EVM address
jq -r '.[] | select(.wallets.eth_evm) | "\(.name): \(.wallets.eth_evm[0])"' data/whitepapers.json

# All papers with a Telegram group, newest era first
jq -r '.[] | select(.socials.telegram) | "\(.dates.likely_year) \(.name) → t.me/\(.socials.telegram[0])"' data/whitepapers.json | sort -r

# Find a person across the whole archive
jq -r '.[] | select(.people.team[]? | test("Terpin")) | .name' data/whitepapers.json

Query the CSV (spreadsheet / pandas):

import pandas as pd
df = pd.read_csv("data/whitepapers.csv")
df[df.likely_year == 2017].sort_values("name")          # the 2017 cohort
df[df.wallets > 0][["name", "socials"]]                  # projects exposing wallets

---

🛠️ How the extraction works

Two small scripts, no third-party Python packages — just pdftotext (Poppler) on your PATH.

Script	Does
tools/extract_metadata.py	Runs pdftotext on every PDF (caching text), then regex-extracts wallets, socials, dates, and role-confirmed names → data/whitepapers.json + .csv
tools/generate_index.py	Renders the JSON into the A–Z data/INDEX.md and the stats in data/STATS.md
tools/build_site.py	Builds the search site index (search-index.json) and commits per-paper plain text to text/ for GitHub-native full-text search
tools/search.py	Offline full-text + metadata search CLI

Re-run after adding PDFs:

python tools/extract_metadata.py     # mines every *.pdf (text cached in tools/.cache_txt/)
python tools/generate_index.py       # rebuilds INDEX.md + STATS.md
python tools/build_site.py           # rebuilds search index + text/

See data/SCHEMA.md for the full data dictionary.

Extraction logic in brief

• Wallets — 0x-prefixed 40-hex for ETH/EVM; bc1… bech32 and 1…/3… base58 for Bitcoin (legacy addresses are kept only when bitcoin/wallet context words co-occur, to cut false positives).
• Socials — per-platform URL patterns; the first handle per platform is surfaced in the index, all of them in the JSON.
• Dates — full dates plus a likely_year = the most frequently cited plausible year in the document.
• Team — names that sit immediately next to a role keyword (Jane Doe, CEO / CTO: John Roe), with a heading/title-word stopword filter. This is the highest-precision signal; loose title-block "authors" are kept separately in the JSON.

---

⚠️ Accuracy & caveats

This is best-effort, heuristic extraction — not hand-curated ground truth. Be aware:

• 222 PDFs are scanned images with no extractable text (has_text: false) — no metadata is mined from them. (OCR would fix this — see Contributing.)
• Wallet addresses may be examples, donation addresses, or citations, not the project's official address. Verify on-chain before trusting any address here.
• Name extraction is regex-based, so expect occasional false positives (a title slipping through) and misses. Treat people.team as leads, not gospel.
• likely_year is a heuristic (most-cited year), which can be off when a paper cites a lot of history.

Found a wrong or missing value? That's a great PR — see below.

---

🤝 Contributing

Stars and forks keep this archive alive — and there's plenty to do:

• ➕ Add a whitepaper — drop the PDF in the repo root, run the two scripts, open a PR.
• 🔍 Improve the extractor — better name detection, OCR for the 222 scanned PDFs, ENS/Solana/Tron address support, founder de-duplication.
• 🧹 Fix metadata — correct a misparsed team or a wrong wallet flag.
• 📈 Add analyses — notebooks mapping addresses, advisor networks, or template clusters.

If this saved you time, drop a ⭐ — it genuinely helps others find it.

---

📜 License & disclaimer

The whitepapers are the property of their respective authors and are collected here for archival, research, and educational purposes. The extraction code (everything in tools/) and the derived datasets (data/) are released for free use.

Not financial advice. Many of these projects are defunct, and some were outright scams. Nothing here is an endorsement. Wallet addresses are extracted automatically and must not be trusted or sent funds without independent on-chain verification.