Hybrid nanostructures are promising for ultrasound-triggered drug delivery and treatment, called sonotheranostics. Structures based on plasmonic nanoparticles for photothermal-induced microbubble inflation for ultrasound imaging exist. However, they have limited therapeutic applications because of short microbubble lifetimes and limited contrast. Photochemistry-based sonotheranostics is an attractive alternative, but building near-infrared (NIR)-responsive echogenic nanostructures for deep tissue applications is challenging because photolysis requires high-energy (UV-visible) photons. Here, we report a photochemistry-based echogenic nanoparticle for in situ NIR-controlled ultrasound imaging and ultrasound-mediated drug delivery. Our nanoparticle has an upconversion nanoparticle core and an organic shell carrying gas generator molecules and drugs. The core converts low-energy NIR photons into ultraviolet emission for photolysis of the gas generator. Carbon dioxide gases generated in the tumor-penetrated nanoparticle inflate into microbubbles for sonotheranostics. Using different NIR laser power allows dual-modal upconversion luminescence planar imaging and cross-sectional ultrasonography. Low-frequency (10 MHz) ultrasound stimulated microbubble collapse, releasing drugs deep inside the tumor through cavitation-induced transport. We believe that the photoechogenic inflatable hierarchical nanostructure approach introduced here can have broad applications for image-guided multimodal theranostics.