API key access now open

Context
compaction
API.

Send long context. Get the useful tokens back before your model call.

Get API key Read docs

70-92% typical token reduction

1 API hosted compaction

0 setup bring your own model

How it works

Keep signal. Drop bulk.

Send context

Pass the query, files, transcripts, or retrieved chunks.

Compact

Radril keeps the tokens that answer the task.

Call your LLM

Use the smaller context with OpenAI, Anthropic, Gemini, or local models.

API

One key. One endpoint.

Usage is measured on source tokens processed by Radril.

curl https://api.radril.com/v1/compact \
  -H "Authorization: Bearer $RADRIL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "What changed in Q4 revenue?",
    "context": "...long document...",
    "target_ratio": 0.10
  }'

Start compacting context today.

See pricing

ContextcompactionAPI.