Scan schematic from image — turn a wiring diagram into a canvas draft

Scan schematic takes an image or PDF of a wiring diagram and drafts the corresponding System Topology on your canvas — instead of clicking every node and edge yourself, the platform reads the diagram and gives you a starting draft to refine.

When to use it

Reach for it whenever the topology exists in some form other than the platform’s own canvas: a paper schematic left by a previous installer, a manufacturer PDF for a specific inverter + battery combo, a hand-drawn on-site sketch, or documentation from a system built before it landed on techs.solar.

If you already have an STX file, use importing-and-exporting-stx instead — it’s exact rather than interpreted.

How to run it

1. Open the plant’s Layout tab.
2. Click the Scan schematic button in the toolbar (right end, teal).
3. Upload the file. Supported formats:
   • JPEG or PNG — photos or scans of paper diagrams
   • PDF — multi-page PDFs are accepted; the reader is instructed to focus on whichever page looks most relevant, though this isn’t a guaranteed mechanism — a single-page PDF is the safer bet
4. Confirm. Processing runs in the background with no progress bar — the plant page refreshes periodically, so the draft topology simply appears on your canvas once it’s ready.

What you get back

The result lands directly on your canvas, already arranged in columns — nothing to place by hand first. Edges get one of four bus types: DC power, AC power, Data, or Control, matching the Layout canvas legend colours.

It tracks its own confidence per item internally, but that doesn’t currently show as a visual flag on the canvas — nothing highlights which nodes it was unsure about. Treat the whole draft as worth double-checking against the original image rather than looking for a specific marker. If you want to tidy the arrangement afterward, Auto-layout will re-flow everything — unconditionally, overwriting manual positioning too, so run it before you fine-tune by hand, not after.

What Scan schematic doesn’t do

• Device linking — it identifies node types, not specific brands or serial numbers. Link each node to a real monitored device on the Devices tab afterward for live data (setting-up-a-monitoring-inverter).
• Full spec extraction — panel wattages, exact inverter models, cable gauges, and voltage levels rarely come through with confidence; fill these in manually.
• Perfect fidelity — it’s a starting point. A messy scan, an unusual layout, or a heavily stylised diagram produces a rougher draft. Always eyeball the result against the original before treating the canvas as authoritative.

When something goes wrong

If a scan can’t be parsed — too blurry, too abstract, an unexpected format — it retries a few times in the background and then gives up with no error shown. If nothing shows up within a few minutes, try a cleaner or higher-resolution scan, a single-page PDF instead of multi-page, redrawing the trickiest section and re-scanning it alone, or just drawing the topology by hand in the Layout Designer if it’s too unusual for the parser.

Where to go from here

• The canvas the scan populates: using-the-layout-designer
• If you have an STX file instead of an image: importing-and-exporting-stx
• Once nodes are placed, connect them to real hardware: setting-up-a-monitoring-inverter