# LLMSlim + OpenAI

> Compress Python application context before an OpenAI SDK request.

LLMSlim is provider-agnostic Python preprocessing. Use RAG provenance for retrieved content.

Category: LLM Provider

## Install

```bash
pip install llmslim
```

## How it fits

1. Retrieve or construct application context.
2. Label provenance accurately and compress with LLMSlim.
3. Send the resulting text to the provider SDK.
4. Apply application-level policy and output validation.

## Example

```python
from llmslim import compress_documents

results = compress_documents(retrieved_docs, query=user_question, target_ratio=0.4)
context = '\n\n'.join(item.compressed_text for item in results)
# Pass context to your OpenAI SDK request.
```

## Deployment

Use LLMSlim in your Python application before constructing the provider request. Measure target workloads before choosing a compression ratio.

## Tips

- Use compress_documents() for retrieved documents; it defaults to RAG provenance.
- Leave system messages unchanged by default in compress_chat_messages().
- Do not treat compression as a complete prompt-injection defense.

## Evidence

| Metric | Uncompressed | Compressed | Impact |
|---|---|---|---|
| v0.4.0 release gate | N/A | 489 tests passed | 90.93% coverage; 375 schemas across 18 catalogs |

## FAQ

### Does LLMSlim guarantee provider output correctness?

No. It reduces context according to its scoring and provenance rules; validate outputs in your application.

## Troubleshooting

- **Unexpected compression result:** Check target_ratio, role labels, and whether the input is long enough to benefit from compression.

Canonical page: https://www.llmslim.app/integrations/openai