Claude Code
Cursor
Codex
OpenCodeFlat monthly plans. Same OpenAI-compatible API.
OpenRouter bills per token and adds a ~5% fee on every credit top-up. Luno runs Claude Opus 4.8, GPT-5.5 and Gemini 3.1 Pro behind one flat plan from $6.99/month — key delivered in about 10 seconds via Telegram.
Luno vs OpenRouter
Side by side, on the things that matter when you ship.
| Feature | Luno | OpenRouter |
|---|---|---|
| Pricing model | Flat plans: $6.99–$39.99/mo | Per-token + ~5% top-up fee |
| Bill predictability | Fixed — same invoice every month | Variable — scales with usage |
| Models available OpenRouter wins on raw catalog size. Luno curates the frontier set most teams actually ship with. | 35+ curated frontier models | 300+ models, widest catalog |
| OpenAI SDK compatible | ||
| Anthropic SDK compatible Point ANTHROPIC_BASE_URL at Luno and the official Anthropic SDK works unchanged. | ||
| Key delivery | ~10s via @LunoCodesBot | Sign-up, then credit top-up |
| Prompt logging | Never — in-memory processing | Off by default, opt-in setting |
| Rate limits | Plan-based, no credit ladder | Scale with credit balance |
| Support channel | Human support in Telegram | Discord + docs |
| Lock-in | None — standard APIs, leave anytime | None — standard OpenAI API |
What actually changes
One bill, zero meter anxiety
Per-token billing punishes heavy weeks. A flat plan means you can let an agent loop overnight without watching a credit balance drain — your invoice is the same number every month.
Key in 10 seconds, no credits dance
No crypto top-ups, no minimum deposits, no fee on every reload. Message @LunoCodesBot on Telegram, pick a plan, paste your key. That is the whole onboarding.
Anthropic-native, not just OpenAI-shaped
Luno speaks both the OpenAI and Anthropic wire formats. Claude Code and the Anthropic SDK connect directly — no translation layer or response remapping.
Switch in two lines
Change the base URL, paste your Luno key. Your SDK, prompts and tooling stay exactly as they are.
# before: https://openrouter.ai/api/v1
export OPENAI_BASE_URL="https://api.luno.codes/v1"
export OPENAI_API_KEY="luno-..."Key delivered in ~10 seconds via @LunoCodesBot on Telegram.
Frequently asked questions
Honest answers on Luno vs OpenRouter.
Yes, for the OpenAI-compatible surface. Swap the base URL to https://api.luno.codes/v1 and replace the key — request and response shapes stay the same. Luno additionally exposes the Anthropic API format, which OpenRouter does not.
No. OpenRouter lists 300+ models; Luno curates 35+. The difference is intent: OpenRouter is a universal router, Luno covers the frontier models — Claude Opus 4.8, GPT-5.5, Gemini 3.1 Pro and their families — that production workloads actually use.
On metered billing, an agent workload that burns 50M tokens a month costs whatever the meter says, plus ~5% on each top-up. On Luno the same month costs $24.99 or $39.99 depending on plan. Light users may pay less per-token; sustained coding or agent workloads almost never do.
No. Plans are $6.99, $12.99, $24.99 or $39.99 per month, billed flat. There are no deposit fees, no per-request surcharges and no expiring credits.
No. Requests are processed in memory and not written to disk. There is no opt-in or opt-out toggle to manage — logging infrastructure for prompt content does not exist in the pipeline.