The complex life of RNA is written in transcript coordinates and chemical marks, yet most genome browsers stay chromosome-oriented and assume a human at the controls, under-serving the data that define RNA biology: ribosome profiling, long-read isoform sequencing, and direct RNA sequencing (DRS). We present VX (https://www.biorxiv.org/content/10.64898/2026.05.17.725790v1.full), a desktop genome and transcriptome viewer (D, GTK 3, OpenGL) that treats genome and transcript coordinates equal. Any transcript FASTA with a transcript-level GTF and a transcriptome-aligned BAM becomes a transcript browser with no reconfiguration, so isoform- and transcript-resolved data are first-class, too. VX renders per-read modification calls, such as m6A, pseudouridine and 5mC from Oxford Nanopore DRS, using MM/ML tags, alongside poly(A) tail-length estimates and per-base mismatch strips. We demonstrate this on human DRS across EEF1A1 and, in yeast, on Saccharomyces cerevisiae aminoacyl-tRNA-seq: dense stacks of short (~60 nt) tRNA reads. Therefore, coverage, sequence, modifications and mismatches read together at base resolution. Crucially, VX exposes its full functionality through an embedded HTTP API and a Model Context Protocol server (39 tools), so scripts and LLM agents observe and drive the view through the same controls as the human, loading data, navigating, running analyses and capturing figures. An integrated framework provides 50+ analyses (signal processing, quantification including RPKM/TPM, peak calling, variant and alignment statistics, cross-track comparison) with a viewport-to-whole-dataset scope hierarchy; outputs are written to disk and re-ingested as new tracks for chained, agent-driven pipelines. VX complements IGV, JBrowse 2 and UCSC by adding native transcript-space handling, modification-aware rendering and an agent-control layer, opening RNA-modification and multi-omic exploration to bench scientists and AI agents alike. Freely available for non-commercial use at https://github.com/Arnaroo/VX.