Free prompt compressor for long LLM context.

Paste RAG context, a support ticket, or an agent trace. Rose 1 returns the compressed context plus a receipt with token savings, ratio, and latency. No signup or API key.

No signupNo API keyLive Rose 1 output

See outcomes

API path

Move the same request body into production.

The public prompt compressor uses the capped demo endpoint. Production traffic uses `/v1/compress` with a scoped bearer key and the same query, input, and compression options.

curl -s https://api.adola.app/v1/demo/compress \
  -H 'content-type: application/json' \
  --data '{
    "model": "rose-1",
    "query": "What should the model answer?",
    "input": "Long RAG, support, or agent context...",
    "compression": { "target_ratio": 0.35 }
  }'