Zero Data Retention

Zero Data Retention is enabled by default for every Serverless request.

You do not need to send a header or change your request. Every request is processed under the same data policy.

How it works

During inference, input data is held in memory and is not written to disk.
We send the output to you and then delete it from memory.
We do not use data you submit to our APIs for training models.
We do not share data you submit to our APIs with third parties.

curl -sS https://api.abhigya.ai/v1/chat/completions \
  -H "Authorization: Bearer $ABHIGYA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen2.5-7b-instruct",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ],
    "max_tokens": 128
  }'

Logs

We do not log the content of your requests. We log only metadata useful for debugging and billing: request ID, cost, and sampling parameters.

Bulk inference

When you submit multiple requests in a single call through our bulk inference APIs, we may need to retain your data for longer, potentially on disk in encrypted form. After inference is complete and we return the results, we delete the data from disk and memory following a short retention period.

Questions

Read the full Data privacy policy or contact contact@abhigya.ai.