Can I use Claude Code with abliteration.ai?

Yes. Claude Code supports Anthropic Messages-compatible gateways. Point ANTHROPIC_BASE_URL at https://api.abliteration.ai and authenticate with your abliteration.ai API key.

Should I use ANTHROPIC_API_KEY or ANTHROPIC_AUTH_TOKEN?

ANTHROPIC_AUTH_TOKEN is still the clearest Claude Code setup because it sends Authorization: Bearer. ANTHROPIC_API_KEY also works when the value is your abliteration.ai ak_ key; it is sent as x-api-key for Anthropic SDK compatibility.

How do I make Claude Code use abliterated-model?

Pick one of two styles. Tier mapping (recommended) sets ANTHROPIC_DEFAULT_OPUS_MODEL, ANTHROPIC_DEFAULT_SONNET_MODEL, and ANTHROPIC_DEFAULT_HAIKU_MODEL to abliterated-model so you can just run claude . Custom model option sets ANTHROPIC_CUSTOM_MODEL_OPTION and you launch with claude --model abliterated-model .

Where should I save the config so it persists?

Either place works. ~/.claude/settings.json scopes the config to Claude Code only. Your shell profile ( ~/.zshrc , ~/.bashrc ) makes the same variables available to any other Anthropic-compatible tool you run from the terminal.

Yes. Claude Code uses Anthropic Messages-style SSE streaming, and abliteration.ai's /v1/messages endpoint supports it natively.

How much does it cost?

With abliterated-model, input is $1 per 1M tokens, cached input is $0.10 per 1M tokens, and output is $3 per 1M tokens. Claude Code sessions can consume credits quickly because they involve many round trips, context reads, and streamed turns.

Docs

Claude Code custom Anthropic gateway

Route Claude Code through abliteration.ai as an Anthropic Messages gateway. Set the base URL and one auth variable, pick a model-selection style, and persist your config via ~/.claude/settings.json or your shell profile.

Updated 2026-08-04

This guide routes Claude Code through abliteration.ai as an Anthropic-compatible gateway. Claude Code is Anthropic's terminal coding agent; abliteration.ai exposes /v1/messages and /v1/messages/count_tokens, so it works as a drop-in backend with environment-variable configuration only.

Setup is three short steps: set the base URL and one auth variable, pick how Claude Code should surface abliterated-model, then persist your config so you don't re-export it in every terminal.

# Quickest path: tier mapping + plain "claude" launch
export ANTHROPIC_BASE_URL="https://api.abliteration.ai"
export ANTHROPIC_AUTH_TOKEN="ak_YOUR_API_KEY"
export ANTHROPIC_DEFAULT_OPUS_MODEL="abliterated-model"
export ANTHROPIC_DEFAULT_SONNET_MODEL="abliterated-model"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="abliterated-model"

claude

Prerequisites

Before you start:

Step 1 — Set the auth variables

Set ANTHROPIC_BASE_URL, then choose one auth variable. ANTHROPIC_AUTH_TOKEN is recommended for Claude Code; ANTHROPIC_API_KEY also works with your ak_ key.

Variable	Value	What it does
`ANTHROPIC_BASE_URL`	`https://api.abliteration.ai`	Routes Claude Code to abliteration.ai instead of api.anthropic.com.
`ANTHROPIC_AUTH_TOKEN`	`ak_YOUR_API_KEY`	Sends your abliteration.ai key as the bearer `Authorization` header. Recommended for Claude Code.
`ANTHROPIC_API_KEY`	`ak_YOUR_API_KEY`	Also supported for Anthropic-compatible tools that send API keys as `x-api-key`.

Step 2 — Pick a model-selection style

Claude Code's model picker is built around Claude model IDs. Two variable sets tell it how to surface abliterated-model. Choose one.

Style	Variables to set	Launch command
A · Tier mapping Recommended. Every default Claude tier (Opus, Sonnet, Haiku) is silently backed by `abliterated-model`, so the normal picker just works.	`ANTHROPIC_DEFAULT_OPUS_MODEL=abliterated-model` `ANTHROPIC_DEFAULT_SONNET_MODEL=abliterated-model` `ANTHROPIC_DEFAULT_HAIKU_MODEL=abliterated-model`	`claude`
B · Custom model option Keeps Claude default models intact and adds `abliterated-model` as an extra picker entry you switch to explicitly.	`ANTHROPIC_CUSTOM_MODEL_OPTION=abliterated-model` `ANTHROPIC_CUSTOM_MODEL_OPTION_NAME=abliterated-model`	`claude --model abliterated-model` (or switch with `/model`)

Step 3 — Persist your config

Exporting variables in every terminal gets old. Pick one place to save them so Claude Code always picks them up.

Location	Where	Use when
Claude Code settings	`~/.claude/settings.json`	You want the config scoped to Claude Code only.
Shell profile	`~/.zshrc` or `~/.bashrc`	You want the same variables available to any Anthropic-compatible tool you run from the terminal.

Example — ~/.claude/settings.json

Claude Code reads an env block from its own settings file. Values here apply whenever you start claude.

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.abliteration.ai",
    "ANTHROPIC_AUTH_TOKEN": "ak_YOUR_API_KEY",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "abliterated-model",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "abliterated-model",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "abliterated-model"
  }
}

Example — shell profile

Append the same values to your shell profile so they export on every terminal launch.

# Append to ~/.zshrc, ~/.bashrc, or ~/.profile
export ANTHROPIC_BASE_URL="https://api.abliteration.ai"
export ANTHROPIC_AUTH_TOKEN="ak_YOUR_API_KEY"
export ANTHROPIC_DEFAULT_OPUS_MODEL="abliterated-model"
export ANTHROPIC_DEFAULT_SONNET_MODEL="abliterated-model"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="abliterated-model"

# Reload
source ~/.zshrc   # or source ~/.bashrc

Verify the connection

Hit /v1/messages directly with curl. If this returns a JSON response, Claude Code will too.

curl -s https://api.abliteration.ai/v1/messages \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $ANTHROPIC_AUTH_TOKEN" \
  -d '{
    "model": "abliterated-model",
    "max_tokens": 64,
    "messages": [{"role": "user", "content": "ping"}]
  }' | python3 -m json.tool

How it works

Claude Code speaks the Anthropic Messages API. Setting ANTHROPIC_BASE_URL redirects that traffic to abliteration.ai. ANTHROPIC_AUTH_TOKEN authenticates as a bearer token; ANTHROPIC_API_KEY is also accepted for ak_ keys via x-api-key.