Welcome! We're excited to offer access to Anthropic's Beta Steering API, which is a beta offering designed to allow developers to adjust internal features of our language models. This is bleeding-edge technology still in the research phase.
We won't be able to "gate in" everyone who requests access, so the more information you can provide on who you are and how you'll use it, the better we'll be able to prioritize.
⚠️ Important: This Beta Steering API is for experimentation only. It is not intended for production use and may changed or be discontinued at any time.
What you get:
- Access to steer a subset of Claude's internal features
- Documentation and example code
- [depending on demand] A Slack channel to discuss with the Anthropic team and other testers
What we ask:
- Share what you build! Just tag @AnthropicAI
- Give us feedback on the API design and your experience
- Remember this is experimental - expect some rough edges
Agreement
By submitting this form, you agree:
- That this preview and your access to the Beta Steering API is subject to Anthropic's Commercial Terms of Service
- Not to use the Beta Steering API in production systems or customer-facing apps
- That Anthropic may revoke access at any time and for any reason
- Not to attempt to use steering to cause harm or bypass Claude's ethical training
- That Anthropic may monitor your interactions with and submissions through the Steering API
- That your participation in this preview and communications in the Slack channel, if invited, will not be confidential
- That Anthropic will not compensate you for your participation in this preview
