Market data and valuations are most useful where your processes run — in your own systems. That is why both pillars of our portfolio are fully machine-readable: the VALUE Market Database via the market data API and VALUE AVM via the AVM API. For AI assistants, we also provide an MCP server, and the REVA valuation environment can be integrated into lending processes.
AVM API — automated valuation as a service
The AVM API delivers value indications as well as market value and mortgage lending value derivations for residential properties and undeveloped land as a REST service (JSON): comparative value, asset value and income value approaches, value ranges and confidence information, plus location and property data relevant to the valuation — from standard land values and location information to market time series. Apart from address and segment, only a few core details are mandatory: living area and year of construction for apartments, additionally plot area and building type for houses, and only the plot area for land. The response arrives within milliseconds.
The interface describes itself: the OpenAPI specification (for customers with access credentials) is available at https://avm-api.value-marktdaten.de/v1/openapi.json, with the corresponding Swagger UI at /v1/api-docs/. The mandatory fields per endpoint are provided by /v1/<endpoint>/specification, for example /v1/indicate/value/specification. The method description, parameter overview and release notes are publicly available as PDFs at /v1/documentation, /v1/parameters and /v1/releasenotes; customers with access credentials can obtain the validation report at /v1/validation. With modelMetrics=true, every valuation also returns the quality of the sub-model used. The quality is independently verified — audited in accordance with IDW PS 880 and with suitability confirmed by an expert opinion within the meaning of § 19(2) BelWertV (German Mortgage Lending Value Ordinance) for the comparative value approach for one- and two-family houses and residential condominiums. Details on approaches, endpoints and evidence can be found on the VALUE AVM page.
Market data API — the market database as a web service
The VALUE market data REST API provides almost all filter options of the market database (location, market and property delimitation) as well as analysis functions as a web service. This allows you to integrate current asking prices, rents and market statistics directly into your applications — the same data basis that underlies VALUE Analyst.
MCP server — VALUE for AI agents
AI assistants and agent systems increasingly speak the Model Context Protocol (MCP) — the open standard through which language models call external tools. Via the VALUE MCP server, an assistant retrieves valuations and location information directly whenever a question calls for it: “What is this apartment worth?” becomes a call to the AVM, and the answer comes with a value range and approach instead of an estimated figure.
The MCP server provides a separate tool for each segment and approach:
- Valuation — full valuation with market value and mortgage lending value for owner-occupied and rental apartments, one- and two-family houses, multi-family houses and land, plus the individual approaches comparative, income and asset value, asking price and energy value.
- Location — micro- and macro-location, standard land value, noise, emissions, climate risks, points of interest, neighbourhood, municipality type, price index and price time series.
- Property characteristics according to NHK — building category, degree of modernisation and fit-out class from individual characteristics, without an address.
Each tool describes its own mandatory inputs. If they are missing, the assistant asks for them instead of inventing values. This gives AI-supported processes the same proven valuation quality as our API customers. The MCP server is available to partners and customers; get in touch if you would like to integrate it into your AI application.
REVA interface — orders and results
In addition to the data interfaces, there is a third route: connecting the REVA valuation environment itself. Property data and inspection results come in this way, and completed valuations flow back into downstream processes — in Europace and Genopace with BaufiSmart without any manual step, within the ongoing loan application.
- Platform case: From the loan application in Europace, REVA opens the valuation directly in the Property tab. Transferred property data is colour-coded as a separate source, and the documents sent along are located in the Review tab. After completion, the platform retrieves the status and documents from REVA.
- Inputs: A valuation can also be created from an AVM JSON or a listing URL. REVA automatically transfers inspection reports from orders placed with VALUE AG into the valuation.
- Outputs: Valuation report, simplified valuation under § 24 BelWertV and compact version as PDF, each generated from the frozen state at completion, plus a JSON output of the data relevant to the valuation for transfer into core banking systems, including via robotic process automation.
Integration, operation and trust
- Documentation: all interfaces documented as OpenAPI specifications, with test environments for integration; new versions appear on the NEXT test environment one week before production, and changes to variables are announced in advance in the release notes.
- Operation in Germany: all services run on our own servers in Germany — without public cloud providers.
- Transparency: valuations with value ranges, confidence information and the underlying data relevant to the valuation — no black box.
- Proven quality: IDW PS 880, § 19(2) BelWertV, annual validation report — see VALUE AVM.
- Proven integration: interfaces to Europace and Genopace with BaufiSmart, among others; more than 200 banks use VALUE values in loan pricing.
A note that also applies to the machine route: the VALUE value indication is a model-based method. It replaces neither experts nor a valuation report — an auditable valuation is produced with an inspection, a review of documents and an assessment by a human being.
Would you like to test the APIs or integrate the MCP server? Contact us — we will set up a test account for you. Terms and conditions can be found under licence models (in German).