Laserlinez converts raster images into line geometry that runs on real machines. The output is built for laser engravers, CNC routers and pen plotters — where path order, node density and closed contours decide whether a file works or fails. The goal is output that is correct for the machine, not just plausible on screen.
CO₂ and diode systems: LightBurn, RDWorks, LaserGRBL. We account for double-burn paths, hesitation at high node density, centerline versus outline strategy and how burn behavior changes across wood, acrylic, anodized aluminium and leather.
DXF output for CAM software such as Fusion 360, VCarve and Aspire: kerf compensation, closed-loop contours for profile cuts and geometry that controllers accept without manual repair.
AxiDraw and compatible plotters: path ordering for single-pen passes, paper behavior under repeated contact and the difference between screen-quality rendering and what actually draws cleanly at 300–600 mm/s.
Readable SVG and DXF for product manuals, technical graphics and CAM handoff — where consistent line geometry matters more than a generic image trace.
ScriptFX is the AI-assisted, non-generative engine behind Laserlinez. Unlike general-purpose vectorizers, it was designed from the start around the constraints of physical output.
It combines deep-learning segmentation — for subject isolation and semantic edge detection — with rule-based computational geometry for path ordering, centerline detection, gap closing and node-density calibration. The result is calibrated for machines, not for browsers.
Version 5 is the current release. Development is continuous and stays close to real production cases, so new workflow requirements feed directly back into the engine. Supported output: SVG, DXF and G-code.
ScriptFX runs automated quality gates that verify output geometry before delivery. We would rather tell you an input needs work than quietly ship a file that fails at the machine.
We test our output formats on real physical machines, not just visually in design software. We keep improving the algorithms and every piece of feedback from production goes into that work — if a file does not run cleanly on your machine, we want to know.
We are a small, focused team. No sales layers, no generic support queue. Your case reaches someone who understands the technical context.
Upload a raster image, choose a conversion preset and preview the line output before you commit to it.