Research Data Consent

Draft, version 2026-09-23. Not yet reviewed by a lawyer.

What you are agreeing to

If you turn on Research Data Consent, you allow syrup to use the content of your conversations with the agent to improve syrup and for research.

Content means: your prompts, the agent's replies, the tools it called and their results, the files it read or changed as part of a session, and signals about the outcome, such as whether you accepted, edited or reverted the agent's changes.

What we do with it

• Study where the agent fails and why, and improve prompts, routing between models, and the tools.
• Build datasets to evaluate models and agent behavior.
• Publish aggregate findings (for example, "model X failed on 12% of edit tasks"). We never publish your code, your prompts, or anything that identifies you.

What we do first

Before content is used for research we run automated redaction for API keys and other secrets, email addresses, and IP-like strings. Research datasets are kept separately from your live account data and are accessible only to people working on syrup under confidentiality obligations.

What this does not include

• Your provider API keys. Never.
• Your account details beyond an internal, non-reversible identifier.
• Selling your data or sharing it with advertisers.

Off by default, and never required

This consent is off until you turn it on. Using syrup does not depend on it. You can switch it off at any time in Settings → Privacy. From that moment we stop using new content, and we remove your existing content from research datasets we control within 30 days. Aggregate findings already published cannot be un-published, but they contain nothing that identifies you.

Legal basis and duration

Your explicit consent (GDPR Article 6(1)(a) and, where applicable, Article 9 does not apply because we do not seek special-category data). Consent lasts until you withdraw it or delete your account. We record the version of this text you agreed to and when.

Questions: sadakhan2002@gmail.com