Changelog
Every change to the Vepler API, in plain language, release by release. New endpoints, improvements, fixes and deprecations.
Subscribe via RSSTitle extractions add alerts for facts no single register entry states, say whether a stated amount is a price, a value or a lease premium, and retry failed extractions.
Sale valuations by location reflect a property's current type and tenure, so a house since converted into flats is valued as a flat.
Sale estimates from predict by location come from a new national valuation model, and sale ranges and confidence scores are calibrated against real sold prices.
Title register entries carry the deed date beside the registration date, and extractions add Property Register rights, the price paid and a free refresh for titles you hold.
Premise matches from address resolve carry an Ordnance Survey classification code, and their position when you ask for it, so one call returns the UPRN, address, point and use.
A Setup page tracks the six steps to your first request, signatures and card payments are confirmed while you wait, and requests are admitted within seconds of a top-up.
publishedDate on listings is optional and may be null when the source publishes no date, and ward and built-up-area crime rows carry geographyFit.
Crime now covers England and Wales with monthly scores, harm rates and trends by area.
Six geography endpoints answer what contains an area, what it contains and where its record came from, with a fit on every row. Land constraint metadata gains two keys.
Three free reads on title deeds: whether a title can be bought, what is pending against it, and what your account has already been charged for it.
meesCompliance gains not_applicable for Scottish certificates and DECs, and address resolve recommends intervention when a sub-building disagrees.
A purchase that HM Land Registry delivers out of band answers 202, and money_state on each money-bearing response says whether credits were taken.
Market status, time on market and lettings price history are built from each advert's own status history, so an agreed sale keeps its price detail.
Each licensed package in the portal's Downloads section shows its release history and when the next release is expected, and its current release is always a full supply.
Address resolve and cleanse hold a postcode disagreement or an unconfirmed flat back for review, so a flat-level input is not auto-accepted against a building-level record.
Every parameter and operation in the API reference is documented for integrators, nine more sections say what they cover, and the property attribute catalogue defines its own fields.
A property query returns up to 10,000 properties in one page, map tiles encode the address fields you ask for, and address suggestions rank around a supplied point.
Census area profiles resolve Scottish geography codes, aggregated crime metrics document their England-only coverage, and listings pin to a dwelling only when the evidence uniquely identifies it.
PAYG customers can request a VAT invoice for a credit top-up self-serve, leasehold titles extract in full without a stated lease term, and every interactive reference example now runs.
Every request body the API reference generates is runnable, leasehold registers with no stated lease term return a complete extraction, and listing locations stay current as new evidence arrives.
Property responses break the address into line_1, line_2, line_3 and post_town, those lines are filterable in property query, and the portal shows per-line-item pricing with your effective rate.
Organisations holding a bulk data licence can now browse and download their licensed datasets directly from a new Downloads section in the portal.
Property query accepts 144 dotted-path filters covering every indexed attribute, marketStatus filter aliases work directly, and area geometry is documented in GeoJSON order with a clear 400.
GET /v1/property/attributes returns every attribute path with its billing tier, live credit price and required permissions, plus the filter and sort whitelists. The dashboard adds a query builder.
With no attributes parameter, property endpoints return locationId, propertyId, slug and address only. Name the blocks you need explicitly; the payload, the tier and the charge then always agree.
The portal's Request Explorer now shows the list-price value a licence absorbed on covered requests, and folds each search typing chain into one row with its settlement.
Every search session settles one flat credit with nothing to complete client-side, messy pasted addresses recover, and AVM valuations and postcode lookups are now priced.
Every endpoint that supports field selection now describes each response field as optional at every nesting level, so any subset of fields you request validates against the published spec.
@vepler/sdk 3.0.0 adds typed clients for heritage, conservation and land constraint designations, school admission areas and search retrieve, plus a new search guide.
Search suggest serves a free candidate set with a search_id, full records come from a charged retrieve step, and responses carry credit headers, rate limits and spend caps.
Property responses carry an explicit countryCode, the address country as a UK nation, the identity scheme behind locationId, and slots for UPRN and Eircode, ready for UK and Ireland.
Search suggest can now return one result per named street, with a representative coordinate, USRN, bounding box and property count, via a new granularity parameter.
The automated valuation model values more property types by location, and returns a specific 400 or 422 when a property genuinely cannot be valued automatically.
Polling a purchased title deed now returns a structured register body: proprietors, charges, restrictions, covenants, notices, class of title and correctly decoded tenure.
New heritage, conservation and land-constraint designation endpoints let you query designations by area, and property responses can now carry heritage and conservation assessments.
The buildings by-point endpoint resolves the building footprints that contain a coordinate, returning them as a standard list response.
A new endpoint returns a school's catchment boundary and indicative last-admitted distance as GeoJSON, and property queries can now filter to the units within a building.
The property vector-tiles endpoint gains a density mode that returns an untruncated per-cell count and average-price grid, and orders individual features by price for stable tiles.
Metric sub-aggregations nested under a bucket on the same sale-history or listings path now return their computed values on the property aggregate endpoint.
The property query, aggregate and tiles endpoints gain two price-reduction filters, and aggregate now returns a 400 when a query names a field it does not support.
POST /v1/flood-risk/point now returns the England Flood Map for Planning zone and NPPF defaults, and POST /v1/property/aggregate exposes its valid field values as a typed enum.
POST /v1/flood-risk/point now reports points outside the modelled flood extents as no modelled risk, out of coverage, or not assessed, rather than one ambiguous value.
POST /v1/flood-risk/point now returns nearby engineered flood defences (type, distance, condition and standard of protection), plus any defended area containing the point.
A new optional addressable_only parameter on search suggest restricts address results to current, addressable properties, excluding historical and superseded records.
A new coordinate flood-risk endpoint returns per-point risk bands and an overall assessment, and schools aggregate blocks now add to the full default record.
The single-school endpoint gains opt-in inspection, identifier and trust blocks, and property search adds verified EPC rating, certificate and fabric filters.
The planning API can return an application's image gallery, directly renderable document files and extracted text, and a summary of its documents, all through new opt-in include options.
More planning applications now carry a site point, so spatial search finds them and returns distances, and decision dates are available to filter and sort on across the planning endpoints.
Search suggest can now bias or restrict autocomplete to a centre point. Pass lat and lng, then choose a soft proximity preference or a hard radius cutoff. Each result returns its distance.
Census flat-file endpoints for tiles, whole-country values and ckmeans breaks resolve the category path parameter and serve data for it.
The census CSV and JSON flat-file endpoints resolve the categoryCode path segment when it carries a .csv or .json suffix.
Census CSV tile and whole-level endpoints return 404 for an unknown category or a geotype the category is not served at, so an invalid request is distinguishable from an empty tile.
The census catalogue and variable endpoints now list only topics that are actually served, report geography levels from real coverage, and tag every variable with a normalised theme.
England & Wales census datasets now serve coarse geography levels (msoa, lad, region, country) across the tile, whole-country and breaks endpoints, not just lsoa.
Planning applications from more local authorities return site coordinates, boundaries and case details, and Scotland census ratios are computed against each topic's own total.
Inside the Greater London LAEI extent, the air-quality point endpoint now resolves NO2, PM10 and PM2.5 from a 20m modelled annual-mean surface instead of the 1km cell plus decay overlay.
The computedAt field on the Prosperity Index endpoint now returns a strict ISO 8601 UTC timestamp, so it passes datetime validation and deserialises cleanly in the SDKs.
New endpoint returns the Vepler Prosperity Index by small-area code, with deciles, dimensions and per-indicator sources. The embedded prosperity object is removed from property responses.
New /v1/census product: datasets, variables, area profiles, custom-area rollups and ONS-format choropleth tiles. Property prosperity scores gain indicators and a classification.
Planning query now supports a near filter to retrieve applications within a metre radius of a point, returned closest-first with a distance field on each result.
New property transactions endpoints, opt-in structured planning extractions via the include parameter, and added air-quality fields for the nearest monitoring station and regulatory-zone charges.
Advanced listing search now returns a canonical tenure.type enum normalised from the raw feed, plus a new tenure.typeRaw field holding the original source string.
New POST /v1/listings/query/advanced endpoint searches aggregated property listings with geographic area filters, a structured field query, sorting, pagination and sparse fieldsets.
Listing refreshes accept source data that repeats a structured feature key, so the features array on every listing endpoint stays current.
New air-quality point endpoint with property-decay overlay, a reworked air-quality query contract (0-100 score, renamed attributes), new schools reference endpoints, and a consistent v2 list envelope on safety routes.
New air-quality query endpoint, expanded safety crime/catalog/geography endpoints, and field-level access control on the buildings API that gates source and structural attributes behind permissions.
New Buildings product exposes building footprints, structure, construction, roof and basement data with lookup, search, batch, spatial and aggregation endpoints.
The Schools API moves to a provider-keyed, query-body shape and gains catchment: per-school polygons, nationwide coverage, and catchment-enriched school queries by UPRN or point.
The areas-within endpoint now returns the documented success/result envelope, with each area's GSS code in the code field instead of id.
Planning applications can now return citation-backed structured extractions via include=extractions. Document type is split into category and detailed sub-type, and address geography now carries the Local Planning Authority.
Added async title deed purchase plus an extraction polling endpoint. Removed the planning policy search endpoint.
The detailed valuation analysis response now exposes per-comparable hedonic adjustments, RICS-aligned confidence decomposition, and price spread reliability metrics.
Compound location queries on the suggest endpoint now boost the specific place name over generic qualifiers like "city centre" or "north", improving result ordering.