Pricing
Mailwoman runs where your code runs, so there is no meter. You either need a release from the AGPL's source-sharing condition or you do not, and the price does not move when your volume does.
Open source — free
AGPL-3.0. That covers the parser, the geocoder, the resolversresolverThe component that converts parsed address components (locality, region, postcode) into coordinates by looking them up in the gazetteer. The resolver ranks candidates by name match, population, and proximity, and returns the best-matching place with its centroid or polygon., the published model weightsmodel weightsThe learned parameters of the neural classifier, shipped as ONNX files in the @mailwoman/neural-weights-* packages. Weights are locale-specific bundles that include the model, tokenizer, and a model-card.json metadata file. and the drop-in API servers. Install it or fork it, and run it in production.
The AGPL attaches one condition. If you distribute software built on Mailwoman, or run it as a network service that other people use, those users are entitled to your corresponding source under the same license. Plenty of teams are fine with that. If yours is not, the next tier is what you want.
Commercial license — $250 a month
Or $2,400 a year, which works out to $200 a month. Per legal entity.
This releases you from the AGPL's source-sharing condition, so you can build closed-source products and services on Mailwoman and keep your source private. One subscription covers one company: every developer, every environment, and every product you ship or host yourself. There are no seats to count and no per-address fees. You also get a direct line to the engineers who built it.
The tier is sized for smaller companies, under 250 employees and under $10M in annual revenue. Past either line, see Enterprise.
Enterprise — from $15,000 a year
Scope agreed with you. This is the tier for address stock that looks nothing like a public corpuscorpusThe BIO-labeled training data used to train Mailwoman's neural classifier. Assembled from real sources (OpenAddresses, National Address Database) and synthetic shards (boundary stress, order variants, negative space). Managed by @mailwoman/corpus.: legacy CRM exports, records typed by field crews, or a countrycountryThe top-level address component (an ISO country). Closed-vocabulary, so it is best handled by a deterministic matcher feeding a proposal rather than a retrained model head. and format the open modelneural classifierThe machine learning model at the core of Mailwoman's parser — a transformer encoder (~30M parameters) trained from scratch to do BIO token classification over addresses. It learns the 'grammar' of address formats; the gazetteer supplies the 'atlas.' has never been graded on. The work is scoped with you, and it has covered a modelneural classifierThe machine learning model at the core of Mailwoman's parser — a transformer encoder (~30M parameters) trained from scratch to do BIO token classification over addresses. It learns the 'grammar' of address formats; the gazetteer supplies the 'atlas.' fine-tuned on your own addresses, an acceptance suite built from your data with the numbers reported to you before anything ships, and support with named contacts and an agreed response time.
Every enterprise agreement includes the commercial license, so the release from the AGPL condition comes with it at any company size. This is also the tier for companies past the commercial license's size fence. The acceptance suite runs on the same discipline as our own published Benchmarks.
If Mailwoman ships inside something you license to other people — a CRM, a billing platform, a device — that is redistribution, and the commercial subscription above does not cover it. Ask about OEM terms: a flat annual fee per product, banded by reach, with no revenue share and no exclusivity.
All prices are in US dollars and exclude VAT and sales taxes.
What we commit to
Your rate is fixed at signup. If the published price rises later, it rises for new subscriptions. Yours stays where it started for as long as you keep it active.
Published releases stay published. Every AGPL release already shipped is licensed to you permanently. That follows from the license itself, and we could not withdraw a grant we have already made.
We do not count your lookups. There is no meter and nothing to true up at renewal. See Cut the per-request bill for the arithmetic against a metered rate as your volume grows.
What we do not commit to
A data refresh cadence. The gazetteergazetteerA geographical index that maps place names and postcodes to real-world coordinates. Mailwoman uses a custom-built Who's On First (WOF) SQLite database as its gazetteer — the 'atlas' half of the grammar/atlas architecture. and layerlayerOne transformer block — attention plus a feed-forward network, with normalization and residual connections — applied to every position. Stacking layers lets the model build up richer representations; Mailwoman's encoder has 6. artifacts are rebuilt when coveragecoverageThe fraction of a population or region for which a data source has real, non-placeholder entries — e.g. 47% rooftop coverage on Texas addresses. Distinct from accuracy on the rows that are present. changes or an ingest bug is fixed, not on a schedule, and no schedule is committed. Each artifact is sealed and dated, so you can always tell which vintage you are running.
A license for the data. The code license and the data license are separate questions. Ours is the code. The reference data comes from open sources with their own terms, and some of those follow the data into your product — read Security and compliance before you ship.