mempool/mempool: self-hosting the mempool.space Bitcoin explorer

Who needs a self-hosted Bitcoin mempool explorer

The project describes itself as a fully-featured mempool visualizer, explorer, and API service, and it is the code running at mempool.space. The stated focus is the transaction fee market that is evolving Bitcoin into a multi-layer ecosystem. That framing matters: this is not a block explorer that happens to show pending transactions. The pending pool is the subject.

The audience splits in two. Most people should use a one-click install method, according to the README, which lists Umbrel, RaspiBlitz, RoninDojo, myNode, StartOS and nix-bitcoin. The second group is developers and people who manage servers, who can use the docker/, backend/, frontend/ and production/ directories. There is also a paid option, Mempool Enterprise, for teams that want the mempool.space team to run the instance on their own infrastructure.

If you only want to look up a transaction occasionally, the hosted site is enough and self-hosting buys you nothing. The self-hosted instance becomes worth the effort when you want the explorer and its API to sit next to your own node, or when you want the interface to keep working independently of someone else's service.

How the repository is put together

The top level separates concerns cleanly. backend/ and frontend/ are the two halves of the application, each with its own manual install instructions oriented for developers. rust/ sits alongside them, which suggests performance-sensitive pieces are not written in the primary language. The primary language reported for the repository is TypeScript, so the bulk of the code is the web application and its server side.

Deployment is handled by three separate directories rather than one: docker/ for container deployment, and production/ for guidance on a more serious instance designed for high performance at scale. That split is a design statement. The Docker path and the production path are not the same path, and the README treats them as different levels of commitment.

At the root there are also nginx.conf, nginx-mempool.conf and http-basic.conf. Those files indicate the project expects to sit behind a reverse proxy and that basic authentication is part of the supported setup. The scripts/ and contributors/ directories, along with .nvmrc, point to a Node version pinned for development. The README does not describe the internal data flow between backend and frontend, so the request path from a browser to the node is not something this article can map.

Installing on a fullnode distro versus Docker

The recommended route is the one-click install, and it is genuinely one step: you install Mempool as an app on a fullnode distro you already run. The README states that no matter which option you pick, you get a fully-sovereign instance up quickly without needing to fiddle with any settings. That claim is only meaningful if the node is already there, since the explorer reads from it.

For developers, the README points at the docker/ directory. The exact compose file and service names are in that directory, not in the README, so read them there rather than guessing. The pattern is a container deployment you configure yourself.

git clone https://github.com/mempool/mempool.git
cd mempool/docker

After cloning, the docker/ directory holds the deployment instructions; follow those files rather than inventing a command line. If you prefer to build from source, the README directs you to backend/ and frontend/ for manual install instructions oriented for developers, and to production/ if you want a high-performance instance rather than a local one.

The first real use is the same in every path: open your instance and look at the mempool view and the fee estimates. The project's own framing is that fees are the point, so the useful first check is whether your instance shows pending transactions and fee rates at all. If it does not, the problem is almost always the connection to the node, not the web interface.

What the repository does not tell you

The README is short and mostly points elsewhere. It does not document rollback, backup, or upgrade procedures for a self-hosted instance, and it does not document the resource requirements of a production deployment beyond calling it high performance at scale. Anyone planning a serious instance should treat production/ as the only authoritative source and expect to read it carefully.

The licence is the larger gap. The repository reports NOASSERTION, which means the automated licence detection could not classify it. The LICENSE and COPYING.md files exist at the root, and the project name carries a registered trademark symbol in the README heading. If you intend to run this commercially, or to fork and redistribute it, the terms you are actually bound by are in those files, and this article cannot tell you what they say. Read them, and take advice if the answer matters to your business.

The wrong-tool case is straightforward. If you do not run a Bitcoin node, or you want an explorer that stays online without you maintaining anything, a one-click install still leaves you with a node to sync and a service to keep alive. The hosted mempool.space exists for exactly that situation.

The hosted service and the enterprise option

The most direct alternative to self-hosting is not another project: it is mempool.space itself, which the README describes as the service this code runs. The difference in approach is operational, not functional. With the hosted site, someone else runs the backend, the frontend, the node connection and the uptime. With a self-hosted instance, you get the same interface and API on your own hardware, and you take on the node, the proxy and the upgrades.

There is a middle option in the same repository: Mempool Enterprise, which the README says is for support on your own production instance, or for having your instance run by the mempool.space team on their own global ISP infrastructure. That is a commercial path rather than a technical one, and it is the answer for a team that wants a dedicated instance without operating it.

A second real alternative is the explorer built into whichever fullnode distro you already use. Those tend to be simpler and narrower: they show your node's view, not a fee-market-focused visualization with a documented API. If your need is a quick transaction lookup next to your node, the distro's own explorer is less work. If your need is the fee market view and an API you can call, the gap is the reason this project exists.

Maintenance, upgrades and the licence question

The repository is not archived and the last push was on 2026-09-24, so the codebase is moving. Recent releases are v3.3.1 and v3.3.0, both dated 2026-04-21, with v3.2.1 before them on 2025-04-14. That release cadence is uneven: two releases on the same day, then a gap of roughly a year before them. Anyone running a production instance should plan for the possibility that fixes arrive in bursts rather than on a schedule, and should watch the release notes rather than assume a regular train.

Upgrade cost depends entirely on which path you chose. On a one-click distro, the distro manages the app lifecycle, so upgrades follow that platform's mechanism. On Docker, you pull and restart. On a manual backend/ and frontend/ install, you own the build and the database, and the README does not describe how to migrate between versions, so that work is undocumented.

On licensing, the honest position is that the repository metadata reports NOASSERTION. The LICENSE and COPYING.md files are present, and the README uses a registered trademark in the project name. If you plan to redistribute, rebrand or offer this as a service, those files are the starting point and not a formality. This is not legal advice; it is a note that the licence metadata alone will not answer the question.

Editorial conclusion

Adopt mempool/mempool if you already run a Bitcoin full node on Umbrel, RaspiBlitz, RoninDojo, myNode, StartOS or nix-bitcoin and want your own explorer and API without touching settings. Do not adopt it if you want a hosted, zero-maintenance service, or if you cannot run and sync a node first. Before committing, check the docker/ directory for the deployment steps, and verify the LICENSE file yourself: the repository reports NOASSERTION, so the actual terms are not something this article can confirm.

Frequently asked questions

Official sources

1. Issues
2. mempool/mempool on GitHub
3. Project website
4. README
5. Releases