Use Claude Code with a custom API endpoint

By the aitokensai editors · Verified 2026-11-02 Sources: Anthropic docs

Claude Code reads its endpoint from two environment variables. Set them once and every session routes through your own key (billed at whatever rate your provider charges).

curl https://aicomp.ai/v1/models \
  -H "Authorization: Bearer sk-your-gateway-key"

Step 1 — get your key and base URL

After signing up, open the dashboard and generate an API key. The dashboard also shows your API base URL. It looks like https://aicomp.ai — without a /v1 on the end, because Claude Code appends /v1/messages itself. You need both values.

Step 2 — set the environment variables

Add these to ~/.zshrc or ~/.bashrc:

export ANTHROPIC_BASE_URL="https://aicomp.ai"
export ANTHROPIC_API_KEY="sk-your-gateway-key"

Then reload your shell:

source ~/.zshrc

If you would rather not touch your profile permanently, prefix a single session instead:

ANTHROPIC_BASE_URL="https://aicomp.ai" \
ANTHROPIC_API_KEY="sk-your-gateway-key" \
claude

Step 3 — verify it actually works

Before you trust it with real work, confirm the endpoint responds:

curl https://aicomp.ai/v1/messages \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-sonnet-5","max_tokens":64,
       "messages":[{"role":"user","content":"say OK"}]}'

A normal completion response means you are routing correctly.

Troubleshooting

Why bother

Claude Code is the single heaviest token consumer most developers run, because it resends accumulated context on every turn. Whatever your monthly bill is today, the per-token rate is the only lever that scales with it. See the Claude Code cost calculator for your own number.

Why changing one URL is enough

An OpenAI-compatible client does not have a provider hardcoded into it. It holds a base URL, and at request time it concatenates that base URL with a resource path — /chat/completions, /models, then sends the same JSON body it would have sent anywhere else, with your key in an Authorization: Bearer header. The base URL is therefore the only thing that decides where the request physically goes.

That is why this change is one line and not a rewrite. Your message array, your tool definitions, your streaming handling and your response parsing are all unchanged, because the wire format is unchanged. What changes is who answers, and what they charge per token. The two values that travel with the new destination are the base URL itself and the API key; the model ID is the third, and it has to be one the new endpoint actually serves, which you can list with a single call to /v1/models rather than guess at.

The practical consequence is that this is reversible. Keep the base URL in configuration rather than inline, and moving between providers is a config change you can undo in a minute — which is also the right way to test it: point a single non-critical call at the new endpoint first, confirm it in the usage log, then move the rest.

What it does not change

Swapping the base URL is narrower than it looks, and the gap between expectation and reality is where most of the frustration comes from.

• Not every endpoint on the surface. Compatibility usually covers /chat/completions and /models. Embeddings, audio, files, fine-tuning and the newer response APIs are a separate question (assume they are absent until you have listed them).
• Not the model's own limits. Context window, output cap, rate limits and moderation behaviour belong to the model and the provider serving it. A shorter context window on the far side shows up as truncation, not as an error.
• Not your timeout and retry semantics. Those live in your client. A gateway adds a hop, so an aggressive timeout that was fine against a direct provider can start failing here, set it explicitly rather than inheriting a default.
• Not your error handling. Status codes mean roughly the same thing but the bodies differ between providers. Anything that parses an error message string is the first thing to break.
• Not the identity of the caller. The endpoint sees one key. If several people or services share it, their spend is indistinguishable in the usage log. Give each one its own key before you need to tell them apart.

What it costs

Rates checked 2026-11-02. Gateway rates move with upstream promotions, verify the current number in your dashboard before committing to a budget.

One month at 20 working days, gateway rates checked 2026-11-02. These are the token volumes, not a quote Your own shape may be larger or smaller.

The three rows are the same three models doing very different work. Chat assistance is dominated by output, the agent loop by input, and bulk processing by neither in particular — which is why the cheapest model changes depending on which row you are on. Model choice matters more here than anywhere else, because agent loops resend context on every turn. The full breakdown is in what Claude Code costs per month.

FAQ

Next

• Claude Sonnet 5 pricing — the usual default for agent coding
• Claude Opus 5 pricing — for the hardest reasoning
• Do the same in Cursor
• Do the same in Cline (the VS Code agent), through its OpenAI Compatible provider
• Do the same in Aider — terminal pair programming on the same endpoint
• Price your own session length in the Claude Code cost calculator
• Eight ways to cut the bill

Get API access