Short answer: Give up now. Conservation of baryon number, lepton number, strangeness, charm, bottomness, and most importantly charge are making your job hard.
Long answer:
I'm ignoring charge conservation issues from time to time; as its no fun if you take that into account
Well, not baryons or leptons. And we have conservation of charge in the way as well.
Baryons
We have conservation of baryon number $B=\frac{1}{3}(\text{number of quarks-number of antiquarks})$, which has opposite values for particle-antiparticle pairs. The only way to convert a baryon to its antibaryon would be to bombard it with a different antibaryon with (negative) twice the baryon number (And twice the negative charge as well). And that would require energy (to create the antibaryon).
Leptons
For leptons, we have conservation of lepton numbers $L_e,L_{\mu},L_{\tau}$. I doubt any reaction exists where a lepton becomes an antilepton, as leptons can only have lepton number $\pm1$, and particle reactions involve two reactants (IIRC). I'm not considering neutrino oscillations here; they make it possible to do stuff like this. The only naturally abundant neutrino is the electron one (maybe not naturally abundant, but easy to generate passively); unfortunately, it's got a very low energy. And the entire neutrino oscillation thing is still debated. Anyways, your neutrino source would be a beta-decaying substance; and there are different,easier ways to get energy from that.
Mesons
It's easier for mesons, though there still are restrictions (conservation of charmness, topness, bottomness). Due to these restrictions, the only conversions possible would be for these mesons (and their antimesons): $u\bar{d} (\pi^\pm),d\bar{t},u\bar{t}$. The last one is anyways in a superposition with its antiparticle, so we get a total of two pairs of particle-antiparticle conversion reactions.
Guage bosons
(I'm not sure if the gauge bosons can do such reactions)
Analysis of feasability of mesons
Anyways, the antimatter=limitless energy is something rather overhyped. Over here, we have two possible candidates. $d\bar{t}$ isn't an everyday particle (Dunno if it's even been synthesized; top quarks are pretty hard to create, and Wikipedia has no data in its list of mesons), and anyways is pretty unstable. You'd have to pump in a bunch of energy to create it, and that defeats the purpose.
Feasibility of pi-meson
$u\bar{d}/\bar{u}d (\pi^\pm)$ is interesting, as it's a common particle in nuclei. But it's bound inside (not exactly--its part of the virtual particle "sea", but that makes it worse), and decays pretty fast. So you'd have to break apart the atom (yes, you'd get energy from that, but not if you separate it into nucleons), "catch" a pi meson, convert it to an antiparticle by means of the "quantum switch", and collide it with another pi meson (alternatively, without the "quantum switch", you can just find an opposite pi-meson). And that will give you a tiny amount of energy as compared to your efforts. You would also need to supply some oppositely-charged particle to conserve charge. Making it more complicated.
Conclusion
So nope, it's not a good energy source. It doesn't work for protons/neutrons/electrons; it only will work for two particles (one more if we consider neutrino oscillations). Neither of them is feasible. Stick to fission-fusion.