OPEN SOURCE · BUILDABLE TODAY · < $1500 RIG

Molecular Synth — a desktop machine that builds
atomically-precise matter by DNA self-assembly

You describe a shape; an AI compiler emits the DNA scaffold + staple oligos to order and the wet-lab recipe; a rig built from repurposed, orderable parts folds it into a real nanostructure — then optionally hardens it into rigid silica.

Honest scope: the output is a DNA-self-assembled nanostructure (~10 nm–1 µm), in ~10¹² copies — not a macroscopic object, and not Drexler diamondoid mechanosynthesis. Every capability cites a demonstrated paper; the bigger “synthesizer” vision is mapped honestly (real-today vs north-star).

A nanofactory you can actually build

DNA origami is the most reproduced result in molecular nanotechnology: a single ~7 kb scaffold strand folded by ~200 short “staple” oligos self-assembles into an arbitrary nanoscale shape, in a single heat-and-cool pot. Molecular Synth turns that into a one-command pipeline + a desktop rig.

< $1500

Rig from orderable parts

Repurposed Peltier thermocycler, DIY gel, optional DIY STM — all linked & priced.

~10–100 nm

Atomic-precision output

Sequence-addressed nanostructures, ~3 nm feature fidelity; ~10¹² copies per reaction.

1 cmd

Shape → recipe

STL/PLY/preset → scaffold + optimized staples + protocol + a 3D oxDNA structure.

6 / 6

Proofs pass

Independent-tool Tm, valid oxDNA structure, simulated control loop, power & fluidics.

How a request becomes matter

shape → A-trail scaffold routing → staple-break yield optimizer (Tm balance · loop-closure · off-target · cross-dimer) → order oligos → fold (90→20 °C in Mg²⁺) → gel-verify → optional silica hardening

What it makes — today

Atomic-precision DNA-origami nanostructures: any shape you describe, folded from a scaffold + optimized staples, in ~10¹² copies per reaction. Everything above this rung is honestly labelled on the vision tab.

A DNA-origami wireframe nanostructure — each edge is a DNA double-helix. The compiler folds shapes like this (~10–100 nm).

🧬 Shapes

Platonic presets, or any STL/PLY/JSON wireframe → a single closed scaffold route + optimized staple set + the wet-lab protocol + a relaxable 3D oxDNA structure.

🔬 Functionalised

Rung-2: handle staples position enzymes / catalysts / nanoparticles at nm-precise sites; cascade placement at a target inter-guest spacing.

🪨 Hardened

Optional sol–gel silica converts the floppy DNA into a rigid SiO₂ replica — a real rung toward durable inorganic nanostructures.

The honest line

The output is a DNA-self-assembled nanostructure (~10 nm–1 µm), not a macroscopic object and not Drexler diamondoid mechanosynthesis. Every capability cites a demonstrated paper; the bigger “matter compiler” ambition stays clearly on the north-star track. We climb one tested rung at a time.

Blueprint — every part, with links & prices

Live from the repo’s single-source bom.json (the same file the validate gate checks). Rig hardware is the one-time build; consumables are per design.

Loading parts…

Build it — the phases

    Full step-by-step (wiring, gel conditions, the rig-validation-first discipline) in the build guide, and the system integration (power, thermal, fluidics, safety) in integration.md.

    The synthesizer thesis — what’s real, what’s north-star

    “Ask for anything, a machine makes it.” A universal replicator isn’t possible today, any price. But the spirit is partly real now — and there’s a real research staircase from DNA toward programmable molecular manufacturing, every rung demonstrated in a lab.

    You ask for…Today?How
    A nanostructure (atomic precision)✓ readythis repo (DNA origami)
    A functionalised / hardened nanostructure✓ readyhandle staples; sol–gel silica
    A custom enzyme / molecule🔬 frontierdirected evolution, DNA-templated synthesis
    Carbon → oxygen / a working phone⭐ north-starnuclear / fabs — not buildable

    The ladder

    DNA origami → breadboard → DNA-templated synthesis → robots → assemblers → general manufacturing. the-ladder.md

    “CRISPR for molecules?”

    Element change is a nuclear wall; molecule building is real chemistry; the tool is engineered enzymes/assemblers. molecule-vs-element.md

    Why it’s real

    A citation per claim, mechanically gated. science.md