Anthropic has some of the best AI models and tools for developers to use. They don’t come cheap though. A lot of folks who use Claude Code don’t mind using other slightly less powerful models in their workflow. With UltraCode_Shim, you can now use Claude Code’s UltraCode with any model you already have access to. For example, you can use GPT 5.5, MiniMax M3, and DeepSeek V4 Pro. Here is what the developers have done to address some of the issues with these kinds of shims:
🔁 Empty turns auto-retry. A backend that returns a turn with no text and no tool call (a transient blip, or a budget-exhausted reasoning turn at high effort) is transparently re-issued. It buffers only until the first real token, so a normal turn adds zero latency and output is never duplicated — and it never retries after real output or a fatal error.
⏱️ A stalled stream can’t freeze the run. If a GPT‑5.5/codex stream opens and then goes silent mid-turn, a bounded idle timeout turns the stall into a quick retry instead of a ~10-minute hang — so one stuck sub-agent no longer freezes an entire multi-agent / dynamic-workflow run.
🛠️ Rejecting a tool call just works. Declining (or skipping) a tool mid-run no longer 400s strict backends like DeepSeek — the proxy repairs the tool-call sequence and synthesizes a stub reply for anything you didn’t answer, including partial parallel calls.
This tool doesn’t edit your global Claude config or credentials. You can find out more about it here.
