> ## Documentation Index
> Fetch the complete documentation index at: https://docs.abliteration.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# FAQ

> Frequently asked questions about abliteration.ai — compatibility, models, retention, and the Policy Gateway.

## Is abliteration.ai OpenAI-compatible?

Yes. Point any OpenAI SDK at `https://api.abliteration.ai/v1`. See [OpenAI compatibility](/api/openai-compatibility).

## Is abliteration.ai uncensored?

Yes. The hosted model is an [abliterated open-weight LLM](/what-is-abliteration) — its refusal direction has been removed at the weight level, so it answers prompts a closed-source model would refuse. No system-prompt jailbreaks needed.

## Does `abliterated-model` refuse requests?

Almost never. The technique surgically removes the refusal subspace from the model's weights, so behavior that comes from refusal training is gone. Capability, instruction-following, and tool use behave like the base model. Edge cases can still occur — that's a residual signal, not policy.

## Can I turn off thinking / reasoning?

Yes — set `"thinking": false` on `/v1/chat/completions` or `/v1/messages`. See [thinking toggle](/capabilities/thinking).

## Is `abliterated-model` a reasoning model?

Yes. Use [`thinking: false`](/capabilities/thinking) to skip reasoning.

## How is abliteration.ai different from OpenAI or Anthropic?

abliteration.ai serves an unrestricted open-weight model through the same OpenAI Chat Completions, OpenAI Responses, and Anthropic Messages request shapes, so any SDK works as a drop-in. Inference is uncensored by default. Governance is opt-in via the [Policy Gateway](/policy-gateway/overview), where you write the rules instead of inheriting someone else's.

## Can I use my existing OpenAI SDK with abliteration.ai?

Yes. Set the SDK's base URL to `https://api.abliteration.ai/v1` and pass your `ak_...` key as the API key — no code changes beyond those two values. Examples for [Python](/integrations/python), [Node / TypeScript](/integrations/node), and [cURL](/quickstart).

## Do you also support the Anthropic Messages API?

Yes. Same base URL — `https://api.abliteration.ai/v1/messages`. The Anthropic SDK works with `api_key=` (sent as `x-api-key`) or `auth_token=` (sent as bearer). See [Anthropic compatibility](/api/anthropic-compatibility).

## Does Claude Code work with abliteration.ai?

Yes. Set `ANTHROPIC_BASE_URL=https://api.abliteration.ai` and `ANTHROPIC_AUTH_TOKEN=ak_...`, then run `claude`. Full setup in the [Claude Code integration](/integrations/claude-code).

## Does Claude Cowork work with abliteration.ai?

Yes. Configure Claude Desktop's third-party inference mode (Developer → Configure third-party inference) with `inferenceProvider: gateway` and `inferenceGatewayBaseUrl: https://api.abliteration.ai`. Full setup in [Claude Cowork integration](/integrations/claude-cowork).

## Can I send images?

Yes, on every API surface. See [images](/capabilities/images).

## Can I send video?

Yes, on the OpenAI Chat Completions surface only. See [video](/capabilities/video).

## Do you support streaming?

Yes, on every API surface. Set `stream: true` (OpenAI) or `stream: true` on the request body (Anthropic). See [streaming](/capabilities/streaming).

## Do you support tool / function calling?

Yes, on every API surface. See [tool calling](/capabilities/tool-calling).

## What model IDs are available?

One: `abliterated-model`. See [models](/models).

## Do you support embeddings?

No. Use a separate embedding provider alongside `abliterated-model` for the LLM.

## Do you support structured outputs (`response_format`)?

No — `response_format` is ignored by the backend. See the [compatibility matrix](/compatibility-matrix) for the full list.

## Do you support web search?

Yes, on all three API surfaces. Three different request shapes — see [web search](/capabilities/web-search).

## Do you support web fetch?

On the Anthropic Messages API only. Not on OpenAI chat completions or Responses. See [web fetch](/capabilities/web-fetch).

## Can I count tokens before sending?

Yes, via `POST /v1/messages/count_tokens`. See [count tokens](/capabilities/count-tokens).

## How do I get an API key?

Sign in to the [console](https://abliteration.ai/console) and create a key. Keys start with `ak_`. See [authentication](/authentication).

## Do you retain my prompts?

Prompt and output content is not retained by default. Operational telemetry (token counts, timestamps, error codes) is kept for billing and reliability. See [security](/policy-gateway/security).

## Do your models have safety filters?

Base inference is unrestricted. The only content always blocked is child sexual content and self-harm. Everything else is opt-in: pass [`flagged_categories`](/capabilities/request-safety-filtering) per request, or use the [Policy Gateway](/policy-gateway/overview) for org-wide rules. If a legitimate use case is blocked, email [help@abliteration.ai](mailto:help@abliteration.ai).

## What's the difference between `/v1/*` and `/policy/*`?

`/v1/*` is a transparent compat surface. `/policy/*` adds project quotas, policy evaluation, policy events, and streaming policy metadata. See [policy endpoints](/api/policy-endpoints).

## Where do I report a bug?

Email [support@abliteration.ai](mailto:support@abliteration.ai) or file an issue in the [console](https://abliteration.ai/console).