My on-chain stack cost me nothing for the first couple of years I traded, and the free version of it still does most of the day-to-day work. Whenever someone asks me which analytics subscription to buy, I ask what question they are trying to answer, and most of the time the answer is sitting in a free tool they already half-know about. The trick is knowing which tool owns which question, because every one of these is excellent at exactly one job and quietly useless outside it. So here is the map, tool by tool, with the point where each free tier stops pulling its weight.
Block explorers are the ground truth
Etherscan, Solscan, Basescan, whatever your chain's equivalent is. The one job an explorer does well is verification. Did this transfer actually happen, what does this wallet actually hold, is this contract verified, what approvals have I granted to contracts I forgot about a year ago. Nothing else answers those questions with the same authority, because the explorer reads the chain directly rather than serving someone's interpretation of it. It is free and it will stay free, and if a paid dashboard ever disagrees with the explorer, the explorer is right.
Where it stops: aggregation. An explorer shows you one address, one transaction, one token at a time. The moment your question becomes what the top holders of a token are doing this month, you are clicking through pages of raw transfers and building the picture by hand, and every paid analytics product on the market lives in exactly that gap. But for a single wallet or a single suspicious transaction, the explorer is the whole tool, and I would trust a beginner who reads Etherscan carefully over an expert who only reads dashboards.
DefiLlama and Dune handle the aggregation
DefiLlama's one job is protocol-level truth, and it does it better than most paid products. TVL by protocol and by chain, fees and revenue, stablecoin supply, unlock schedules, funding rounds, even a running list of hacks. No account, no paywall, open source. My most used page is fees and revenue, because it answers the question that kills most DeFi narratives: does this protocol actually earn anything, or is the yield just token emissions recycled through a prettier dashboard. Where it stops is the wallet level. DefiLlama tells you what protocols are doing in aggregate and nothing about what any specific holder is doing.
Dune's one job, on the free tier, is that someone has almost certainly already built the dashboard you want. DEX volume by chain, stablecoin flows, holder behavior for a specific token, whatever narrative is running, search for it and a public dashboard usually exists. Browsing is free. Where the free tier stops is freshness and custom work. Free query credits are limited, heavy queries wait behind paying users, and a public dashboard is only as current as the last time someone refreshed it, which can be days or weeks. My rule before trading on anything from Dune is to check the last refresh timestamp, and if real money hangs on the number, to skim the underlying SQL, which is public. Dashboards inherit the assumptions of whoever wrote them, and some of those assumptions are wrong in ways the chart will never show you.
Glassnode and CryptoQuant, the free slice
Glassnode's free tier does one thing I still use: long-horizon context on the majors. Core metrics like active addresses and transaction counts at daily resolution, going back years. If your question is shaped like whether the current market resembles past cycle tops or bottoms, the free tier covers a surprising amount of it. Where it stops: resolution and depth. The cohort and entity-adjusted metrics that professionals actually argue about sit behind the paywall, and daily granularity is close to useless for anything you plan to trade within the week. Treat free Glassnode as a monthly context check rather than a timing tool.
CryptoQuant's free slice is about exchange flows. Reserves held on exchanges, inflow spikes that historically precede sell pressure, plus a community feed where analysts publish free interpretations of the same charts. Where it stops: lookback and alerting. The free window of history is short, so you can see that inflows are elevated but you cannot easily line that up against past extremes, and you have to check manually because alerts are a paid feature. That last part generalizes across this whole list, because the underlying data is mostly public and what you pay for, almost everywhere, is history depth and a machine watching the feed so you do not have to.
Whale feeds, and wiring it all together
Public whale trackers, the Whale Alert style feeds on X and Telegram, plus Arkham's free entity labels and DeBank for reading any wallet's full portfolio. The one job: knowing within minutes that a large transfer happened, and roughly which entity moved it. Where they stop is the why, and this is where beginners lose money. Exchanges shuffle funds between their own hot and cold wallets constantly, and every one of those internal moves reads like a whale depositing tens of thousands of ETH to sell if you only see the headline. I have watched people short what turned out to be routine custodial housekeeping. Before reacting to any whale alert, check the labels on both sides of the transfer, and if either address belongs to the same exchange's own infrastructure, it is almost certainly noise.
Here is the zero-cost workflow I actually run when a token starts moving and I want to know if it is real:
- Explorer first. Is the contract verified, and how concentrated is the holder list. A handful of unlabeled wallets holding most of the supply changes everything downstream.
- DefiLlama. Are fees, revenue, or TVL moving with the price, or is this pure price action with nothing underneath.
- Dune. Search the token, check the refresh date, and look for real user counts rather than raw volume, which is easy to fake.
- Exchange flows. Is the token flowing onto exchanges, which usually means distribution, or off them, which usually means accumulation.
- Whale feeds. Is any labeled foundation, fund, or early investor wallet on the move, keeping the exchange-internal caveat in mind.
That loop takes maybe fifteen minutes and costs nothing, and it has filtered out most of the trades I am glad I never took. The rule of thumb for when to start paying is simple. Upgrade when you catch yourself running the same manual check every single day, when you need an alert instead of a habit, or when the free history window cuts off exactly where your question begins. We reached the same conclusion building Blockcircle, where most of the whale and disclosure data we track is public and the real product is the filtering and alerting layered on top of it. Run the free stack for a month, write down which checks you repeat, and pay for only those. Most people who try this end up paying for a lot less than they expected.