Give it the sequence you want to build and it returns the oligonucleotides that assemble it. The construct is tiled by oligos that alternate between the two strands and overlap their neighbours; in the first cycles those overlaps anneal and the polymerase extends each oligo along its partner, so the fragments grow towards full length. End primers are then added for a second round of cycling, which amplifies the finished construct away from everything incomplete.
What the design actually optimises is the melting temperature of the junction overlaps — one cycling program has to serve every junction at once, so an even Tm across them matters more than even oligo lengths. Method: polymerase cycling assembly; Stemmer et al. (Gene 1995). Long constructs are split into blocks that are assembled separately and joined — for that last step see Gibson Assembly.
One sequence, 5′→3′, with or without a FASTA header — the finished construct as you want it to end up. Anything that is not A/C/G/T is dropped and reported: an ambiguity code cannot be built as a single oligo set.
Mask before placing junctions
Blocks
Export: use Copy / Download above each report, or click inside a result area → Ctrl+A → Ctrl+C.