Full-spec Claude API
Stable, reliable, ready to use
Same models and capabilities as official, no degradation. Available worldwide, pay per use.
Trusted by 10+ companies, 30+ API resellers and 7,000+ users
FAQ
What is derouter.ai?
derouter.ai provides Claude and GPT APIs with the same versions and capabilities as official. One account, one API key for every model — pay per use, no subscription, Claude from 1/4 of the official list price.
How does the cost compare with official APIs?
Claude models cost about 1/4 of the official list price, and GPT models as little as 10%. Casual users typically spend under $1 a month; heavy developers save tens of dollars monthly. Pricing is fixed and transparent — no hidden multipliers.
Which clients and SDKs are supported?
derouter.ai is a drop-in replacement for the OpenAI and Anthropic APIs — just change the base URL. For Claude Code, set ANTHROPIC_BASE_URL=https://api.derouter.ai/proxy; for Codex CLI, configure a custom provider as shown in the docs; any standard OpenAI SDK or compatible client works. Cursor is not supported at the moment.
Where is it available? What payment methods are supported?
Available worldwide with direct access — no VPN required. Sign up with email and pay with crypto or bank card.
How is it different from OpenRouter?
Both offer OpenAI-compatible APIs. The key difference is pricing: Claude and GPT on OpenRouter are billed at roughly official list prices, while derouter.ai uses fixed discounted pricing — Claude at about 1/4 of list and GPT from 10% — plus flexible payment options such as crypto.
Which models are available?
Currently available: Claude Fable 5, Claude Opus 4.8 / 4.7 / 4.6, Claude Sonnet 5 / 4.6, Claude Haiku 4.5, GPT-5.5, GPT-5.4 / 5.4 Mini, plus GPT Image 2 for image generation. One unified API, one key.
Is there a free trial or a minimum top-up?
The minimum top-up is just $10. No monthly fees, your balance never expires, no commitments. You are only charged for the tokens you actually use.
Is the service reliable? Is my data safe?
In stable operation since 2025, available 24/7, and your account is unaffected by regional restrictions — no risk of bans. Conversation content is used only to fulfil your requests and stays private. Optional TOTP two-factor authentication protects your account.
How do I get started?
Sign up with your email in 30 seconds, top up from $10, then call the API. No KYC, no phone number, no credit card needed.
Do you modify my requests or inject a system prompt?
No. Your requests are forwarded verbatim — no prompt rewriting, no injected instructions, no model swapping, no post-processing. The only addition is a ~140-character standard client identity header; your own system prompt and tool definitions pass through unchanged. The identity part is served from a 1-hour prompt cache automatically, so its cost amortizes to near zero in ongoing conversations.
A third-party authenticity checker shows one failed item — is the model degraded?
No. Requests reach Anthropic through the official Claude Code client channel and responses come back byte-for-byte, so you get the 100% official full-spec model. A few checks (such as forced structured output) show unsupported because the Claude Code channel itself lacks that capability — a channel difference, not a counterfeit or degraded model. Official fields like thinking signatures are present in every response and can be verified yourself.
