Framework integration
DSPy compression integration
DSPy optimizes LLM prompts automatically. SuperCompress adds a compression layer that reduces token costs before DSPy's prompts reach the model.
DSPy integration
import dspy
from supercompress import Compressor
comp = Compressor()
class CompressedModule(dspy.Module):
def __init__(self, signature):
super().__init__()
self.signature = signature
def forward(self, **kwargs):
# Compress long context inputs before DSPy processes them
for key, val in kwargs.items():
if isinstance(val, str) and len(val) > 1000:
kwargs[key] = comp.compress(val, str(self.signature)).compressed_text
return super().forward(**kwargs)
Frequently asked questions
Does compression affect DSPy optimization?
No. DSPy optimizes the prompt structure, not the input context. Compression only reduces input size.
Can I use it with DSPy's teleprompters?
Yes. Compression works before the teleprompter's optimized prompt is assembled.
Build with less context
Put compression in front of your next LLM call.
Use the hosted API or run SuperCompress locally. Keep the evidence, drop the token waste, and measure the savings before it reaches your model.