AI Generation · DERpass v1.0

From plain text
to passport.

Describe your PV system in plain language — an installation report, a product invoice, a WhatsApp message from the installer — and an AI model will convert it into a valid .derpass file. No manual schema browsing required.

Three steps from
raw data to passport.

01

Copy the prompt template

Below you'll find a ready-made system prompt. It gives the AI context about DERpass, links to the schema, and sets clear output rules. Copy it once — reuse it for every system.

02

Paste your unstructured data

Add your raw source at the end of the prompt: a commissioning report, an equipment list, a data-sheet scan, or even a rough description you type yourself. The AI extracts what it can and leaves unknown fields out.

03

Validate and open in viewer

Copy the JSON output into a .derpass file. Drop it into the DERpass Viewer to check it renders correctly, or run it against the JSON Schema with any validator.

Unstructured input,
structured output.

A technician's handover note converted to a complete DER Passport. The AI maps every piece of data it can recognise to the correct schema field.

Input — technician's note
handover-note.txt
Anlage Musterweg 12, 80331 München Eigentümer: Petra Müller Inbetriebnahme: 17.03.2025 Module: 24x Trina Solar Vertex S+ TSM-430 Wechselrichter: Fronius Symo GEN24 10.0 Plus Seriennummer: 34591827463 AC-Leistung: 10 kW Dachfläche Süd, Neigung 30°, Azimut 185° Kein Speicher MaStR-Nummer: SEE948273649102 EAN (Zählpunkt): DE0023781000000000000000003847 Vergütung: EEG-Einspeisevergütung 8,20 Ct/kWh Netzbetreiber: Bayernwerk Netz GmbH Installation durch: Sonnenkraft Bayern GmbH Zertifikat: ISFH-2024-9938 Erstprüfung bestanden, keine Mängel.
Output — generated .derpass
musterweg-12.derpass

Copy, paste,
describe your system.

Works with Claude, ChatGPT, Gemini, or any capable language model. Replace the placeholder at the end with your own data.

derpass-ai-prompt.txt
You are a DERpass expert. DERpass is an open JSON standard for describing photovoltaic (PV) systems and other distributed energy resources (DER). Schema reference: https://derpass.org/core/v1.0/schema.json Your task is to convert the unstructured data I provide into a valid .derpass file that conforms to DERpass Core v1.0. Rules: - Output only raw JSON. No markdown fences, no explanation text. - Only include fields for which you have data. Never invent or guess values. - Required top-level fields: derpass_version ("1.0"), jurisdiction, meta, commissioning. - meta must contain at least: system_id (generate a short slug if not given), name, owner.name, and location.country. - commissioning must contain at least: commissioning_date, dc_capacity_kwp, ac_capacity_kw, and module_count. - Use ISO 8601 dates (YYYY-MM-DD). - Capacities in kWp / kW / kWh. Power per module in Wp. - jurisdiction.primary uses ISO 3166-1 alpha-2 (e.g. "DE", "NL", "AT"). - For German systems include national.DE with mastr_number if available. - If the source mentions a feed-in tariff, populate compensation.entries[0] with scheme "fit" and the tariff rate in eur_per_kwh. - Installer details go into commissioning.installer (name, cert). - Place all timestamped events in the events array with type, timestamp, actor.name, and summary. --- SOURCE DATA --- [Paste your commissioning report, invoice, equipment list, or system description here. Plain text, tables, or copied PDF content all work.]

Getting better results.

Give more, get more

The AI only populates fields it has data for. An invoice that includes serial numbers, metering point IDs, and installer certificates will produce a much richer passport than a rough description.

Validate the output

Run the result through the Viewer or a JSON Schema validator before treating it as authoritative. AI models can occasionally mis-map a field or hallucinate a value.

Iterate in chat

If a field is missing or wrong, correct it in the same conversation: "Change the feed-in tariff to 0.0792" or "Add a maintenance event on 2025-09-10." The model updates the JSON in place.

+

Multiple source documents

Paste several documents in one go — commissioning report, then the equipment datasheet, then the insurance policy. The model merges them into a single passport, resolving any conflicts.

National profiles

Mention the country upfront. For German systems the model will include national.DE fields such as mastr_number and balkonkraftwerk. For Dutch systems it adds national.NL with EAN-17 and SDE++ fields.

Portfolio generation

For many systems at once, ask the model to produce a JSON array of DER Passports — one per system — then split the array into individual .derpass files with a short script.

Any format your
installer gives you.

AI models handle all of these without any preprocessing on your side.

PDF

Commissioning reports, datasheets, grid connection approvals — copy the text content and paste it directly into the prompt.

CSV / Excel

Component lists, yield tables, meter readings. Paste the rows as-is — tabs and commas are both fine.

Email / chat

A forward from your installer or a WhatsApp screenshot description. Informal language is fine — the model extracts what it can.

Monitoring portal export

SMA Sunny Portal, Fronius Solar.web, SolarEdge — paste the system summary text or the downloaded report.

Free-form description

No documents at all? Just describe the system in plain language. The model will ask clarifying questions or leave unknown fields out.

Existing JSON / XML

Already have data in another format — SunSpec, EEBUS, or a proprietary API response? The model maps it to DERpass fields.

Open the result in the viewer.

Once you have a .derpass file, drop it into the viewer to get a rendered, printable document — no upload, no account.

Open Viewer See Examples