LotusAI

Flat-rate coding API

Coding intelligencethat never runs dry.

One key for Claude Code, Cursor, Cline and every Anthropic-format or OpenAI-format tool. Generous five-hour windows that refill on their own, so a long agentic run never dies halfway through.

Get access Read the docs
  • 14 models
  • 1M context
  • Drop-in endpoint
  • OpenAI-compatible

5-hour window

refills on its own

live
window usageresets every 5h

one

key

any

tools

5h

limit

zsh
$ export ANTHROPIC_BASE_URL=https://api.lotusai.pro$ export ANTHROPIC_AUTH_TOKEN=sk-lotus-...$ claude> connected. window refilling

Works where you work

Change one base URL. Keep your whole setup.

Claude CodeCursorClineContinueZedRoo CodeOpenCode
Anthropic Messages APIOpenAI /v1/chat/completionsStreaming SSETool useAny Anthropic-format toolAny OpenAI-format tool

Why LotusAI

Built for people who ship every day.

Everything a key gets you, with none of the metering anxiety.

One key for every tool

The same key works in Claude Code, Cursor, Cline, Continue, Zed, Roo Code, OpenCode and anything else that speaks the Anthropic or OpenAI format.

Five-hour windows that refill

No monthly cliff. A generous rolling window starts on your first request and resets on its own every five hours.

Streaming and tool use

Server-sent streaming, tool calls and long agentic loops behave exactly as your tools expect them to.

OpenAI-compatible endpoint

Prefer the OpenAI shape? Point at /v1/chat/completions with the same key and carry on.

Live key status, no login

Paste a key on the public status page to see its window, usage and reset time. Nothing to sign in to.

Check a key

Partner-issued keys, human support

Keys come from authorised partners who know the product, and real people answer on Telegram and email.

How it works

Three steps. A few minutes.

01

Get a key from a partner

Keys are issued by authorised partners. Message one, receive your sk-lotus key, done.

02

Point your tool at the endpoint

Set the base URL and your key in Claude Code, Cursor or any Anthropic-format or OpenAI-format tool. One config change.

03

Build

Stream, call tools, run long sessions. Your five-hour window refills on its own while you work.

Get access Setup guide

The lineup, 14 models

Every model. One key.

Up to 1M of context. Pick a model per request with the model field, or pin one in your tool's config and forget about it.

FrontierNew

Claude Fable 5.1

the new heavyweight

The most capable model on the list: demanding reasoning and long-horizon agentic work. Knows the world up to June 2026, the most recent cutoff of anything we serve.

Context

0M

1,000,000 tokens

Access

One

key for every model

Questions

Quick answers.

The short version. The full FAQ covers setup, limits and troubleshooting.

Read the full FAQ

  • Which tools does a key work with?

    Claude Code, Cursor, Cline, Continue, Zed, Roo Code, OpenCode and any tool that speaks the Anthropic Messages API or the OpenAI chat completions format.

  • What happens when my window runs out?

    Requests pause until the five-hour window resets, then everything resumes on its own. The public key status page shows the exact reset time.

  • How do I get a key?

    Keys are issued by authorised partners. Message a partner on Telegram or email support and you will have a key shortly.

    Get access

Your tools.One key. Never dry.

Get a key from a partner, point your editor at the endpoint, and build without watching a meter.

Get accessRead the docs
Get access Docs