Integrations

Works with the traces you already have.

No new SDK and nothing in your request path: rightmodeler reads the traces your agent already emits, replays them through cheaper models, and proves which swaps are safe.

Reads your traces

Autodetected and folded into one per-step schema.

Claude CodeRun the rightmodeler skill on ~/.claude/projects: it replays every session step through cheaper models, judges each against what you accepted, and reports the safe swaps.
Codex CLIPoint rightmodeler at ~/.codex/sessions and it replays each step of your Codex CLI rollouts through cheaper models, judging every output against what you accepted.
LangSmithExport a LangSmith run tree and rightmodeler replays each step through cheaper models, judges every output against what you shipped, and reports the safe swaps.
the OpenAI SDKPoint rightmodeler at your OpenAI SDK request log and it replays every call through cheaper models, judges each output against what you shipped, and reports the safe swaps.
LangfuseExport Langfuse observations to disk and rightmodeler replays each step through cheaper models, judging every candidate against the output you already shipped.
BraintrustFetch Braintrust span rows and rightmodeler replays each step through cheaper models, judges every candidate against the output you shipped, and reports the safe swaps.
PhoenixExport OpenInference spans from Phoenix and rightmodeler replays each step on cheaper models, judges every output against what you shipped, and reports proven swaps.
OpenTelemetryPoint rightmodeler at OTLP span exports carrying gen_ai.* attributes. It replays each step on cheaper models and proves which swaps hold up against shipped output.

Replays through

Where the candidate calls actually run.

OpenRouterSet one OPENROUTER_API_KEY and rightmodeler shortlists cheaper candidates from the live catalog, replays your steps at temperature 0, and prices every swap with usage.cost.
LiteLLMrightmodeler re-runs your real pipeline with a LiteLLM proxy injecting the cheaper candidate, tools fire, loops run, and the report shows whether the swap survived.

On the roadmap

Named integrations arriving with Crucible.

Vercel AI GatewayNo native Vercel AI Gateway adapter yet: audit AI SDK steps today where your telemetry emits gen_ai.* spans, and join the Crucible waitlist for native support.
RelatedHow it worksReduce LLM costsGlossary