Burnless – open protocol that cut my API bill 16x on my heaviest dev day (github.com)

Burnless has introduced an open protocol that significantly reduces API costs for multi-agent workflows by offering a vendor-agnostic orchestration layer. By allowing developers to utilize various language models (LLMs) from different providers—like Claude, GPT, or local models—Burnless shifts the cost structure from O(N²) to O(N), thereby cutting expenses on multi-turn interactions. This is achieved through a shared cached system prompt and efficient history management, where only short capsules are stored and reused, leading to substantial savings on token costs throughout conversations. The implications for the AI and machine learning community are profound. Burnless not only allows for flexibility in selecting models but also enforces user-defined routing and cost rules, creating a more controlled and predictable financial landscape for developers. The system leverages multiple layers of compression, minimizing input costs while sticking to a mathematical approach rather than marketing claims. Initial benchmarks show that developers can witness savings up to 16 times, making Burnless an attractive solution for those looking to optimize their API usage without sacrificing performance.