Predio wraps Spain's Catastro in JSON for agents that cannot live with SOAP

Predio launched a pay-per-call JSON and MCP service that wraps Spain's public Catastro cadastral data into machine-readable APIs for proptech teams, real estate operators and AI agents, bypassing the official SOAP/XML interface. The startup offers REST and MCP endpoints covering 78.9 million properties in Spain's common territory, with machine-readable pricing and stale-response caching designed for agent workflows.

Predio https://prediohq.com/?ref=runtimewire is turning Spain's public Catastro into a pay-per-call JSON and MCP service for proptech teams, real estate operators and AI agents, with a sharper bet than a basic API wrapper: the company is packaging old government web services as agent infrastructure. Predio's public site is written in the first person, but the builder's identity is not disclosed on the pages reviewed by RuntimeWire. That absence matters because the product has the shape of a founder-led developer tool rather than a broad enterprise data vendor: a narrow data problem, an opinionated API contract, self-serve pricing, machine-readable docs and an early adjacent product aimed at land due diligence. The pitch is simple. Spain's cadastral data is public, but Predio says the official interface remains hard for machines to consume because developers have to deal with SOAP/XML, inconsistent error handling and schemas oriented toward human use. The official Sede Electronica del Catastro web-services page https://www.catastro.hacienda.gob.es/ayuda/servicios web.htm?ref=runtimewire still documents Catastro web services and says the system is in a technology migration, with older services expected to be replaced over time. Predio is building on top of that reality instead of waiting for the government interface to look like a modern developer platform. Predio says its API covers Spain's common-territory cadastre, excluding Pais Vasco and Navarra, which have their own regional cadastres. Its site cites 78.9 million properties in the common territory as of January 2026, split between roughly 40.2 million urban properties and 38.7 million rural properties. Predio presents those figures as Catastro baseline counts, and it is explicit that it is a non-official service using data from the Direccion General del Catastro. The wrapper is the wedge Predio's first product resolves and enriches a Spanish property through three inputs: cadastral reference, coordinates or address. The normalized output covers fields such as property class, use, built area, year built, address and either construction or cultivation breakdowns. The company says it uses the same output shape for urban and rural properties, which is the kind of detail that matters when the customer is an agent or backend workflow rather than a person clicking through an official viewer. The REST interface sits under /v1, while Predio links an MCP endpoint at https://api.prediohq.com/mcp https://api.prediohq.com/mcp?ref=runtimewire . Predio lists three MCP tools in its llms.txt https://prediohq.com/llms.txt?ref=runtimewire and OpenAPI: consultar inmueble catastro , consultar inmueble por coordenadas and consultar inmueble por direccion . Predio also publishes an OpenAPI spec https://prediohq.com/openapi.yaml?ref=runtimewire , machine-readable pricing JSON https://prediohq.com/pricing.json?ref=runtimewire and a login flow for API keys, usage and balance. That stack explains the timing. A developer wrapper for Catastro could have existed as a REST business years ago. The MCP layer changes the buyer story. Predio is not only selling fewer XML headaches to developers; it is selling a tool surface that an AI agent can discover, call and chain during a task such as property valuation, site screening or portfolio enrichment. The machine-readable pricing page is part of the same bet: agents and developer tools need to understand costs before they call. Predio's error contract is also a product decision. Its docs say errors return JSON in a stable { "error": { "code", "message" } } shape with standard HTTP codes, and that business errors such as 404 and 422 do not consume credits. Cache hits and serve-stale responses do. The stale-response policy is a tell. Predio says it can return the last cached official value marked stale if the source falls over. That makes sense for agent workflows, where a failed tool call can derail an entire run. It also creates a trust burden: Predio has to make freshness obvious enough that a valuation tool, due diligence workflow or real estate CRM does not treat stale official data as live official data. Pricing starts as a developer-tool motion Predio's pricing is usage based. Its pricing JSON https://prediohq.com/pricing.json?ref=runtimewire says one credit equals one successful response. The free tier gives 250 credits per month. Paid prepaid packs start at EUR 5 for 2,500 credits, then EUR 9 for 5,000, EUR 29 for 25,000 and EUR 99 for 150,000. The largest listed pack comes out to EUR 0.00066 per successful call. Predio says prepaid checkout runs through Stripe and lists X402 as upcoming. Those are starter numbers. The service is selling reliability and data normalization on top of a public source, rather than exclusive data ownership. The question for a buyer is how much uptime, normalization and agent compatibility are worth compared with consuming the official service directly or using an existing data platform. Predio is also testing a second product line. ParcelGuard https://prediohq.com/parcelguard/?ref=runtimewire is in early access and aims to return parcel constraints by cadastral reference, including flood exposure, Red Natura 2000, coastal-domain restrictions and livestock-route layers, with source and date for each layer. A sample ParcelGuard report https://prediohq.com/parcelguard/ejemplo/?ref=runtimewire shows how Predio wants to move from property lookup into due diligence screening. That is the more valuable wedge if Predio can execute it. A clean Catastro wrapper saves developers time. A constraint report that stitches together Catastro geometry, flood layers, protected-area data, coastal restrictions and other public sources moves closer to a due diligence product that can command report-level pricing. Predio is entering a crowded wrapper market Predio should not be treated as the only company trying to modernize Catastro access. CatastroAPI https://catastro-api.es/?ref=runtimewire markets a REST API for Spain's cadastre with JSON responses, endpoints for provinces, municipalities, streets, properties and coordinates, plus a 14-day free trial. Goolzoom https://www.goolzoom.com/en/api/?ref=runtimewire offers a broader Spanish geographic-data API built around the Cadastre, Land Registry, INE and other sources. DataPrem https://dataprem.com/en/?ref=runtimewire is positioning Spanish public data as tools for agents and teams, including Catastro, BORME and tenders. Parse.bot also lists a Catastro Spain API https://parse.bot/marketplace/03640b79-4739-4c60-907d-f22bc9f5662a/catastro-minhap-es-api?ref=runtimewire in its marketplace. That competitive field narrows the claim Predio can credibly make. REST and JSON alone are table stakes. Predio's sharper angle is the combination of a versioned JSON contract, MCP, llms.txt, machine-readable pricing, JSON error semantics and serve-stale behavior, all around one narrowly defined Spanish property-data problem. The company is effectively saying that old public data systems need an agent-facing operations layer: schema, pricing, errors, cache semantics, provenance and tool discovery. The funding story is absent. Predio does not disclose investors, valuation, revenue, headcount or customer names on the reviewed pages. It also does not disclose the founder's name, prior work or company entity on the public materials supplied. That keeps the story product-led for now. Predio has a visible API surface and a clear pricing model, but the durability of the business will depend on whether developers and real estate teams pay for operational polish around data they can technically reach without Predio. For a small tool, that is a reasonable place to start. The official Catastro remains the source of truth. Predio's job is to make the source usable in the places the official interface was not designed for: automated valuation, agentic workflows, bulk enrichment, due diligence pre-screening and backend systems that cannot parse ambiguity for a living.