About
About BitcoinBagger
A free site that tracks the Bitcoin price, who holds Bitcoin and the money moving through Bitcoin ETFs, built entirely from public data that refreshes every few minutes.
The site
What BitcoinBagger is
BitcoinBagger brings together figures that are otherwise spread across several websites: the Bitcoin price, how much Bitcoin the spot ETFs and listed companies hold, the daily ETF flows and a handful of market signals such as Fear & Greed and Bitcoin dominance. We put them on one site so they can be read side by side, with the time each figure was last checked printed next to it.
Everything comes from free public sources, which we name below, and the site doesn’t ask you to sign up or pay. That keeps it simple, although it also means our figures are only as good and as current as those sources, which is why we show where every number came from.
Bitcoin data sources
Where every figure comes from
Each source has its own timetable, because some figures move every minute and others change once a day at most.
| Data | Source | How often we check |
|---|---|---|
| Bitcoin price, market value and 24-hour change | CoinGecko, with Binance and then DefiLlama as backups | Every 5 minutes |
| Bitcoin dominance and total crypto market value | CoinGecko, plus one dominance reading a day that we record ourselves | Every 10 minutes |
| Daily price history for the past year | CoinGecko | Every 6 hours |
| Bitcoin held by companies, ETFs, miners and governments | Bitbo (BitcoinTreasuries.com), with CoinGecko’s company list as a backup and cross-check | Every hour |
| US spot ETF daily flows | Bitbo | Every hour |
| US share and ETF prices (delayed) | Nasdaq | Every 5 minutes while the US market is open, hourly otherwise |
| Market capitalisation and fund assets | Nasdaq | Every 12 hours |
| Daily price history for five watchlist stocks | Nasdaq | Every 6 hours |
| Weekly Bitcoin closes for the DCA backtest | Kraken | Once a day |
| Perpetual futures funding rate and open interest | Deribit | Every 15 minutes |
| Blocks, hashrate, fees and halving estimate | mempool.space | Every 5 minutes |
| Crypto Fear & Greed Index | alternative.me | Every hour |
| Pound to dollar reference rate | European Central Bank | Every 6 hours |
| Crypto app rankings | Apple App Store charts | Every 6 hours |
| News headlines | CoinDesk, Cointelegraph, Bitcoin Magazine, Decrypt and The Block, through their RSS feeds | Every 15 minutes |
Refresh cycle
How the data reaches the page
The path from a source to the number you read, and the delays along the way.
-
Step 1
A five-minute snapshot
A scheduled job runs every five minutes, asks each source that’s due for fresh figures and saves the results as one snapshot, which every page reads from.
-
Step 2
Failures keep the last good value
If a source doesn’t answer, we keep its last good figure with the time it was fetched and try again within fifteen minutes, so you see an older number clearly dated and never a blank or a guess.
-
Step 3
A short cache in front
Netlify’s CDN keeps each live page for up to five minutes, and may serve that copy for up to ten minutes longer while it fetches a new one, which keeps the site quick at the cost of figures that can be a few minutes behind the snapshot.
-
Step 4
A backstop for stale data
If a page is asked for when the snapshot is more than 20 minutes old, it refreshes the quicker sources itself before it renders, and falls back to the saved snapshot if that takes longer than eight seconds.
AI summaries
How the AI summaries are written
The short briefs on the home page and the notes beside the largest holders are written by Claude, Anthropic’s AI model, from the live figures. A job runs every fifteen minutes and sends the model a fixed set of numbers from the latest snapshot, with instructions to describe what happened in plain British English and to avoid predictions or advice.
- Labelled every time - each block shows whether it came from the AI or from our rules, the model name and the time it was written.
- A rule-based fallback - when the AI is unavailable, or a brief is more than 90 minutes old, the page shows an automatic summary built by fixed rules from the same figures.
- Notes on holders - notes on individual funds and companies are kept for up to eight hours before they’re replaced by rule-based text.
- A backup model - if Claude can’t be reached and an OpenAI model has been configured, that model is used instead, and the label says so.
Limits
What BitcoinBagger doesn’t do
- No advice - we aren’t financial advisers and aren’t regulated by the FCA, so we describe what the data shows and never tell anyone what to buy, sell or hold. Our disclaimer sets this out in full.
- No business - the site is run for education and entertainment, it sells nothing and it carries no adverts, sponsored posts or affiliate links.
- No accounts - there’s nothing to sign up for, so we don’t hold logins, portfolios or watchlists.
- No trading - you can’t buy, sell or store Bitcoin or any share through the site, and we don’t hold anyone’s money.
- No cookies or tracking - the site’s own code sets no cookies and runs no analytics or adverts, as our cookie policy explains.
Questions
Questions about BitcoinBagger
How often is the data on BitcoinBagger updated?
A scheduled job runs every five minutes and checks which sources are due, and each source has its own timetable, which is set out in the table on this page. The Bitcoin price is checked every five minutes, dominance every ten, treasury holdings and ETF flows every hour, and Nasdaq share prices every five minutes while the US market is open but only hourly when it’s shut, because the price doesn’t move then. Every figure carries the time it was last checked, so if something looks old you can see straight away how old it is.
Why might a figure on the site be out of date?
Most often it’s because the source itself hasn’t updated yet. Company holdings only change when a company files or announces a purchase, ETF flow figures arrive only after each US trading day has closed, and Nasdaq quotes are delayed. Less often a source fails to respond, and when that happens we keep showing the last good value with its time and try again within fifteen minutes, which is why some pages show an amber note saying a source isn’t updating. If a figure still looks wrong after allowing for that, we’d like to hear about it through the data error form.
Who writes the AI summaries, and can I trust them?
Claude, Anthropic’s AI model, writes them from a fixed set of figures we pull out of the latest snapshot, and it’s told to describe what happened without predictions or advice. We label every block with the model and the time it was written, so you can tell an AI brief from our automatic summaries, which are built by simple rules from the same numbers. We’d still treat them as a quick read of the figures and not as analysis, because a model can phrase things more confidently than the data deserves, and the tables next to each summary are the place to check anything that matters to you.
Does BitcoinBagger cost anything, or need an account?
No to both. Every page is free to read and there’s nothing to sign up for, which also means we don’t hold accounts, passwords or portfolios for anyone. The only personal details that reach us are the ones you choose to send, through the data error form or by email, and our privacy policy covers what we do with them.
