Built API-first, and it shows
Most explorers are websites that later exposed their index as an API, and the result is usually an API shaped by the website's needs rather than a developer's. BlockCypher went the other way.
The consequences are visible throughout. The documentation is organised around tasks rather than endpoints. Error responses are informative. The data model is consistent across chains, so code written for Bitcoin works for Dogecoin by changing a path segment.
It also explains the web interface, which is plainly an afterthought and has been for years. That is a reasonable allocation of effort for a company whose customers are integrating rather than browsing — but it does mean that if you arrived looking for an explorer, you arrived at the wrong product.
Webhooks, which almost nobody else offers
If you need to know when a payment arrives, the naive approach is polling — ask every thirty seconds whether anything changed. It works, it wastes everyone's resources, and it scales badly.
BlockCypher lets you register a webhook against an address or transaction and receive a POST when something happens: a transaction appears in the mempool, reaches a confirmation threshold, or is double-spent. That last one is genuinely valuable and rarely offered.
The alternatives are running your own node with ZeroMQ notifications, or mempool.space's WebSocket, which pushes updates but requires you to hold a connection. For a server-side integration that wants a callback rather than a socket, BlockCypher's model is the simplest thing that works.
Transaction construction, and its trade-off
BlockCypher will build an unsigned transaction for you — select inputs, calculate change, estimate the fee — returning something for you to sign locally and send back for broadcast.
That is a real convenience. Coin selection is fiddly, fee estimation is easy to get wrong, and change handling is where people leak privacy or lose money. Having it done correctly by a service is attractive.
It is also a meaningful trust decision, and worth being explicit about. The service learns your addresses, chooses which inputs to spend, and sees the transaction before the network does. Your keys never leave your machine — the signing is yours — but the metadata exposure is substantial, and the input selection affects your privacy in ways you have delegated.
For a production wallet, construct transactions locally. For a prototype, an internal tool or a low-stakes integration, this is a reasonable shortcut that saves real work.
The free tier, honestly
BlockCypher's free tier has per-second and per-hour limits, with a token required above them. It is enough for development and testing and it is not enough for anything with users.
That is a fair commercial position, and it sits differently from the free tiers elsewhere in this directory. Blockstream and mempool.space serve read-only data at no charge with no key, funded as public goods. BlockCypher is a business selling infrastructure, and its free tier is a trial rather than a service.
So the sensible split is: read-only data from the keyless APIs, and BlockCypher when you need webhooks or the write path — the things nobody gives away. Our API comparison lays out where each sits.
What our reference tests showed
We tested BlockCypher as a developer product rather than as a website, since that is what it is.
The read endpoints behaved correctly across Bitcoin, Litecoin and Dogecoin. Address balances, transaction details and block data all reconciled against Blockstream and Blockchair. The consistent schema across chains is real — the same parsing code worked on all three without modification, which is not true of stitching together three chain-specific APIs.
The webhook test is the one that mattered. We registered a hook against a test address and it fired promptly on the unconfirmed transaction and again at the confirmation threshold we set. The payload was well structured and included enough to act on without a follow-up request, which is a detail that separates a usable webhook from a notification that triggers three more API calls.
Rate limiting arrived exactly where the documentation said it would, which is worth noting because it is not universal. Several services in this directory throttle earlier than their published figures imply. BlockCypher's limits were as advertised, and the 429 responses were informative rather than opaque.
The transaction construction endpoint produced a valid unsigned transaction with sensible input selection and a reasonable fee estimate. We signed it locally and broadcast it successfully.
The web interface, tested alongside, did work with JavaScript disabled — a point in its favour that its appearance does not suggest. But it showed less on every lookup than the free alternatives, and there is no version of our recommendation where someone browses here rather than building here.