Unrestricted AI.You govern.
A policy gateway and developer-controlled LLM API for teams that need full control over what their models can do.
Stream a completion. No signup.
Talk to abliterated-model from your browser. Conversation stays in this device, never on a server.
Control plane
included.
Default public APIs ship a single model behavior. abliteration.ai ships the governance layer as part of the product—so you decide what gets through.
Policy Gateway
Define policy-as-code guardrails with rewrite, redact, escalate, or refuse outcomes. Manage per-project keys and quotas, test with shadow mode and canary rollouts.
Unrestricted LLM models
Developer-controlled, less-restricted models without provider-side refusals. Built for security research, fictional writing, medical research, and other high-control workloads.
Synthetic data generation
Generate training data, fine-tuning pairs, eval sets, and JSONL-ready outputs through the same OpenAI-compatible API used for live inference.
Audit & compliance exports
Export every policy decision to Splunk, Datadog, Elastic, S3, and Azure Monitor with structured metadata for internal investigations and audits.
Sign up.Get a key.Ship.
Swap the
base URL.
Full /v1/chat/completionsparity with the OpenAI API. Keep your SDKs, your tools, your code—change one line.
Works with all major SDKs. Set ANTHROPIC_BASE_URL for Claude Code, switch wire_api for Codex, or just point your OpenAI client.
Simple,
predictable pricing.
Works with
your stack.
Use your existing AI coding tools and SDKs. Stream audit logs straight into your SIEM with structured metadata for compliance.
Unrestricted,
not ungoverned.
You control refusals with policy-as-code. We keep the audit trail you need—without storing prompts or outputs by default.
Zero data retention
Prompts and outputs processed transiently, never used for training.
Policy-as-code
Define your own rules. Rewrite, redact, escalate, or refuse.
Decision metadata
Every policy outcome is logged with policy ID, reason code, and tags.
SIEM exports
Stream to Splunk, Datadog, Elastic, S3, and Azure Monitor.
One line
to migrate.
Drop-in replacement for OpenAI and Anthropic clients. Keep your SDKs, your code, your tooling. Change the base URL.
OpenAI wire format
Same request/response shape. Change the base URL and go.
Streaming responses
Server-sent events supported on /v1/chat/completions.
Vision input
Send images alongside text; metered as input tokens.
Synthetic data
JSONL-ready outputs for training, fine-tuning, and evals.
curl https://api.abliteration.ai/v1/chat/completions \
-H "Authorization: Bearer $ABLIT_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "abliterated-model",
"messages": [
{ "role": "user", "content": "Hello!" }
],
"stream": true
}'Pay for
what you use.
Builder
Individual developers seeking access to unrestricted AI models.
/ month — 6,666,667 tokens (13,334 credits)
- Unrestricted models included
- Billed on combined input + output tokens
- OpenAI-compatible /v1 endpoints
- Vision / image understanding
Team
Teams requiring priority throughput for unrestricted models.
/ month — 16,666,667 tokens (33,334 credits)
- Everything in Builder
- Priority throughput
- Higher concurrency limits
- Email support
Policy Gateway
Governance starts at $600/mo
/ month — Control / Advanced / Enterprise
- Policy-as-code rules with reason codes
- Simulation + shadow / canary rollouts
- Audit logs with change history
- Cross-app enforcement at Advanced+
- Org-wide governance at Enterprise
Ship with a
model that ships.
Swap your base URL and go. 5 free requests to try it, prepaid credits if you like it, Policy Gateway when you scale.
5 free requests · no phone verification · OpenAI-compatible
