abliteration.ai
AI API and policy gateway compatible with OpenAI and Anthropic
About abliteration.ai
abliteration.ai provides an AI API compatible with OpenAI Chat Completions and Anthropic Messages, for red teams, security and trust, synthetic data, machine learning research, and defense and government workflows. The platform batch-generates preference pairs, evaluation rows, and classifier examples, and applies allow, refuse, rewrite, redact, or escalate processing through a policy-as-code gateway. Results can stream reason codes to Splunk, Datadog, Elastic, or your own S3.
Best for
Best use cases
Generate preference datasets
Build evaluation and classifier examples
Integrate existing AI SDKs
Enforce enterprise policies and auditing
Features
Multi-protocol API compatibility
Compatible with OpenAI Chat Completions API and Anthropic Messages API for integration with existing SDKs.
Training data generation
Generates preference pairs, evaluation rows, and classifier examples by mode, with schema validation.
Policy-as-code gateway
Configure custom policies for model requests and return allow, refuse, rewrite, redact, or escalate results.
SIEM audit output
Stream request-processing reason codes to Splunk, Datadog, Elastic, or your own S3.
Pros
- Compatible with OpenAI Chat Completions and Anthropic Messages API
- Supports batch generation of preference pairs, evaluation rows, and classifier examples
- Processes requests under policy with five outcomes and returns reason codes
- Can send logs to Splunk, Datadog, Elastic, or S3
Pricing Plans
Developer
- 2.5% usage discount
- Pay-as-you-go token usage
- OpenAI-compatible /v1/chat/completions
- Web search
- API keys with project limits
- Auto-reload supported
- Prepaid credits never expire
* Pricing information is for reference only. Please refer to the latest announcement on abliteration.ai's official website.