https://secantusdb.com/2026-05-05T00:00:00+01:00The SQLite of document databases2026-05-05T00:00:00+01:002026-05-05T00:00:00+01:00Joe Drumgooletag:secantusdb.com,2026-05-05:/blog/hello-world/

<p>Welcome to the SecantusDB blog — what the project is, what it isn't, and where it's heading next.</p>

<p>Welcome to the <strong>SecantusDB</strong> blog. SecantusDB is a real MongoDB server written in Python, backed by the same WiredTiger engine MongoDB ships with. It speaks the MongoDB wire protocol on a real TCP socket, so <code>pymongo</code>, <code>mongo-go-driver</code>, <code>mongosh</code>, and <code>mongodump</code> all connect unchanged. Within single-node scope, your driver can't tell us apart from <code>mongod</code>.</p> <h2 id="what-its-for">What it's for</h2> <ul> <li><strong>Tests.</strong> Spin up an embedded server in a fixture, run your suite, tear it down. No Docker, no ports to manage, parallel-test friendly.</li> <li><strong>Dev loops.</strong> Local development without standing up a <code>mongod</code> process. Default storage is on-disk WiredTiger, so your data survives a restart.</li> <li><strong>Single-node prototypes.</strong> Small services that don't need a real cluster but want a real document database with real durability.</li> </ul> <h2 id="what-it-isnt">What it isn't</h2> <p>SecantusDB is <strong>not</strong> a replacement for production MongoDB. It does single-node only by design. Replica sets, sharding, real cluster topology, and anything that depends on multi-node behaviour are out of scope. We advertise ourselves as a single-node <code>secantus</code> replica-set primary in the <code>hello</code> reply so <code>pymongo</code>'s topology machinery accepts change streams, but the topology is fictional &mdash; there is no other member.</p> <h2 id="status-alpha">Status: alpha</h2> <p>We're in early development. The Python API surface (CLI flags, public class signatures) is still settling and may shift between point releases. The on-disk format is WiredTiger's, and our schema on top of it (collection / index / oplog tables) has been stable across releases. There's no migration tool yet, so don't put production data here.</p> <h2 id="whats-next">What's next</h2> <p>More driver-conformance coverage, performance work on the aggregation pipeline, and richer index acceleration (multi-field sort, collation). Follow along on <a href="https://github.com/jdrumgoole/SecantusDB">GitHub</a> or read the <a href="https://secantusdb.readthedocs.io/">docs</a>.</p>