QuiverAI’s Arrow 2 can reconstruct a PNG or JPEG as an editable SVG, aiming to replace part of the manual work involved in tracing old logos and other raster graphics. Rather than producing a vector-look image alone, the system generates SVG files built from editable shapes and paths.
A public test by Ahmet Doğan used the model to recreate the Manchester United crest, illustrating the type of logo-conversion task Arrow 2 is targeting. QuiverAI says its technology is focused on AI-generated vector graphics rather than general-purpose image generation.
The company’s approach is based on generating the drawing structure itself, including shapes, paths and control points that can be adjusted after export. In 2026, QuiverAI raised an $8.3 million seed round led by Andreessen Horowitz, according to the reporting dossier.
Editable paths matter more than a convincing preview
The key question for designers is whether the exported file can be refined efficiently. A logo can look accurate in a preview but still be difficult to use if it contains excessive nodes, overlapping objects or paths that are cumbersome to modify.
The useful question is whether the resulting paths are organized enough that a designer can refine them.
Om Kamath
That makes final-asset time a more meaningful measure than the first generated image. If a file delivers most of a logo and requires only a few curve corrections, it may remove a substantial share of the tracing work. The result still needs to be inspected in a vector editor before it can be considered production-ready.
Arrow 2 outputs SVG elements that can be adjusted after generation, including the ability to:
- Recolor shapes and fills.
- Move vector elements.
- Correct paths and geometry.
For more complicated images, QuiverAI offers Arrow 2 Telos, a version positioned around more careful preservation of detail and geometry. Arrow 2 can also generate SVGs from text prompts or reference images, extending its scope beyond raster-to-vector conversion to icons, UI illustrations, diagrams and web assets.
The technology is available through the QuiverAI app and is integrated with Paper, where users can generate a vector element on a canvas and continue editing its paths, fills and strokes. QuiverAI also provides an API for developers building the technology into their own workflows.
Adobe remains a major reference point for professional vector workflows, including its Firefly vector-generation tools. QuiverAI’s potential differentiator is its focus on the SVG as the primary output. The remaining test is whether its generated files consistently need only light cleanup rather than a full redraw.
