Lithium (Li) ions are central to the energy storing functionality of rechargeable batteries1. Present technology relies on sophisticated Li-inclusive electrode materials to provide Li ions and exactingly protect them to ensure a decent lifetime2. Li-deficient materials are thus excluded from battery design, and the battery fails when active Li ions are consumed3. Our study breaks this limit by means of a cell-level Li supply strategy. This involves externally adding an organic Li salt into an assembled cell, which decomposes during cell formation, liberating Li ions and expelling organic ligands as gases. This non-invasive and rapid process preserves cell integrity without necessitating disassembly. We leveraged machine learning to discover such functional salts and identified lithium trifluoromethanesulfinate (LiSO2CF3) with optimal electrochemical activity, potential, product formation, electrolyte solubility and specific capacity. As a proof-of-concept, we demonstrated a 3.0 V, 1,192 Wh kg−1 Li-free cathode, chromium oxide, in the anode-less cell, as well as an organic sulfurized polyacrylonitrile cathode incorporated in a 388 Wh kg−1 pouch cell with a 440-cycle life. These systems exhibit improved energy density, enhanced sustainability and reduced cost compared with conventional Li-ion batteries. Furthermore, the lifetime of commercial LiFePO4 batteries was extended by at least an order of magnitude. With repeated external Li supplies, a commercial graphite|LiFePO4 cell displayed a capacity retention of 96.0% after 11,818 cycles.
