Edge-deployed systems such as autonomous robots, AR/XR devices, and emergency-response handhelds require accurate indoor topological representations, yet existing sensor-based and expert-curated mapping methods are impractical for crowdsourced, resource-constrained deployment. We present Lang2Graph, an experimental framework for indoor topological graph inference from natural-language navigational instructions that isolates four governing factors: instruction structure, metadata clarity, prompting strategy, and model size and reasoning capability. We also propose the Independent Prompt Executor (IPE), a prompting strategy that decomposes graph construction into independent per-instruction reasoning steps to reduce error propagation. Across synthetic and augmented real-world benchmarks, structured instructions, clear metadata, and IPE substantially improve graph inference quality and point toward viable on-device edge deployment.