Context Optimizer

Compresses your prompts. -40% tokens. Same output quality.

How it works

Before forwarding your request to the LLM, GateCtr analyzes and compresses the prompt:

• Removes redundant whitespace and filler phrases

• Condenses verbose instructions without changing intent

• Trims conversation history to the most relevant turns

• Preserves all semantic meaning and context

Average reduction: -40% tokens. Output quality is maintained.

Enable

Context Optimizer is enabled by default on Pro plans and above.

To control it per request:

Node.js

```typescript const response = await client.complete({ model: 'gpt-4o', messages, gatectr: { optimize: true }, });

console.log(response.gatectr.tokens_saved); // e.g. 312 </div><div data-gb-custom-block data-tag="tab" data-title='Python'>python response = client.complete( model="gpt-4o", messages=messages, gatectr={"optimize": True}, )

print(response.gatectr["tokens_saved"]) # e.g. 312 </div><div data-gb-custom-block data-tag="tab" data-title='cURL'>bash curl https://api.gatectr.com/v1/complete -H "Authorization: Bearer $GATECTR_API_KEY" -d '{ "model": "gpt-4o", "messages": [...], "gatectr": { "optimize": true } }'

## Response fields

```json
"gatectr": {
  "optimized": true,
  "original_tokens": 800,
  "tokens_saved": 320,
  "compression_ratio": 0.40
}

Disable for a specific request

gatectr: { optimize: false }

Useful for requests where prompt precision is critical (e.g. structured output, code generation with exact formatting).

Available on

Pro plan and above.