Hey - for pricing, we're going usage based on two metrics : amount of third-party connections and volume of data transformed (for chunking / embedding). Ps: This will evolve in the next months probably!