Built Ondine on top of LiteLLM, a batch DataFrame processing SDK, would love an integration listing

[ptimizeroracle]

Hey LiteLLM team 👋

Built an open-source SDK called Ondine that uses LiteLLM under the hood for provider routing. It wraps batch LLM processing around a Pandas or Polars DataFrame you give it a prompt, a model, and output columns; it handles retries, checkpointing, cost caps, and structured output via Pydantic.

LiteLLM is what makes the "same code, any provider" part work. Users swap model="gpt-5.4-mini" for model="ollama/qwen3.5" or model="anthropic/claude-sonnet-4-6" and nothing else changes. That's entirely thanks to your abstraction.

from ondine import QuickPipeline

result = QuickPipeline(
    source="products.parquet",
    prompt="Extract brand, category, sentiment from: {description}",
    output_columns=["brand", "category", "sentiment"],
    model="gpt-5.4-mini",  # or any LiteLLM-supported model
    budget_limit=25.00,
).run()

Repo: https://github.com/ptimizeroracle/ondine
Website: https://ondine.dev

Ask: Would it be worth adding Ondine to your integrations / projects-using-LiteLLM list? Happy to open a PR against the docs if that's the preferred path.

Also, genuinely, thanks. LiteLLM saved me weeks of provider-specific adapter work.