What about a wood fired baking oven. There are versions derived from the rocket stove that are quite efficient
The general need here, I think, is food-grade controllable heat delivery. This utility is often combined into an oven and range, and allows for heat processing of agricultural and food products. Perhaps keeping that in mind is worthwhile when it comes time to design this machine.
I also don't think this will be a one-size-fits-all solution. Other machines will enable the manufacture of different types of ovens & ranges, but a general design goal is getting the most useful food processing out of a given amount of fuel. Traditional earthen ovens cook different dishes at different temperatures as the oven cools from its initial firing.
It'd be interesting to combine a given design with a heat transfer model to be able to come up with a temperature over time profile and estimate how much cooking could be done. One possible start to creating that sort of model would involve combining FreeCAD Python scripting to handle a design's geometry, in addition to leveraging the existing work of the ht library (https://github.com/CalebBell/ht), a component of Caleb Bell's Chemical Engineering Design Library (ChEDL), an MIT-licensed group of Python libraries.