A viewer shows you the drawing. A parser gives your software the data.

A CAD viewer is a rendering engine. It reads a DWG or DXF file, interprets the geometry, and draws it on screen. The output is pixels — a visual representation for a person to look at. Examples range from desktop applications like AutoCAD and DraftSight to browser-based viewers like Autodesk Viewer.

A CAD parser is a data extraction engine. It reads the same file and produces structured output — JSON objects describing each entity, its coordinates, its layer membership, and its properties. The output is machine-readable data for software to process. CADLens is a hosted REST API that does exactly this, and optionally adds a rendered PNG so you also have a visual.

A common pattern in early-stage products is: open the DWG in a viewer, take a screenshot, display the PNG. This works for basic display, but it breaks the moment you need more:

• No units. A PNG pixel has no real-world dimension. You cannot recover the length of a wall or the diameter of a circle from a raster image without manual re-digitisation.
• No layer information. Layer names, colors, and per-layer entity counts are gone as soon as the drawing is flattened into an image.
• No entity data. You cannot count the number of doors, filter structural members, or identify text annotations from pixels.
• No automation path. Every analysis step would require human eyes and manual input, which does not scale.

If your product needs to measure, classify, filter, or transform CAD content, you need a parser — not a screenshot.

Use a CAD viewer when:

• Engineers or end-users need to interact with drawings — pan, zoom, annotate, mark-up.
• You are building a design review or collaboration tool.
• Your users need to query drawings manually using a UI.

Use a CAD parser like CADLens when:

• Your software needs to read and process drawing data without human intervention.
• You need to extract measurements, entity counts, or layer content programmatically.
• You are building quotation, estimation, BOM extraction, or floor-plan processing features.
• You want to display a non-interactive thumbnail of the drawing alongside structured data.

Use both when your product offers engineers an interactive view (via a viewer) and also runs automated processing in the background (via CADLens).

CADLens is a parser, but it also renders a PNG preview from the same parse run. A single call to POST /v1/parse returns a job_id. When the job completes, the result includes:

• vectorJson — every entity with coordinates, layer, color, and bounding box.
• layersJson — layer names, colorHex, and entityCount per layer.
• metadata — drawing units and bounding box ({ width, height }).
• imageUrl — a signed URL to a rendered PNG preview, available immediately alongside the JSON.

You display the PNG in your UI and process the JSON in your backend — from one API call, with no additional rendering infrastructure required.