Skip to main content

All the world's AI, on one URL.

AICU API is an API platform for AI creators that lets you use the world's leading AI models with a single API endpoint.

Get started from beautiful TTSGet a FREE API credit

The best endpoint for creators in the AI creative era.

→ Have AiCuty (ElevenLabs-compatible TTS) read your long-form text!

High-quality TTS from AiCuty, the idol created by humans and AI, is here! Use ElevenLabs-v2-compatible multilingual TTS — plus v3 Enhanced, which reads long-form Japanese correctly with rich emotional expression — right now.

Budgetable

Prepaid. What you charge is your budget — when it runs out, it stops. Errors are free, cache hits are free, and an idle app costs nothing.

OpenAI-compatible

Swap the base URL and API key — existing OpenAI SDK, OpenCode and Cline code just works.

Characters and voices

Keep your character's face consistent with GPT-Image2 reference images, and give them a voice with ElevenLabs character presets.

Coding AI

Put advanced coding assistants like OpenCode to work right now, for free.

Make you AI creator together.

Using Claude Code, Codex or another coding agent? Paste the block below as-is and it will wire your project up. Then just keep going: "add a read-aloud button", "turn this article into an MP3".

Integrate the AICU API (api.aicu.ai) into this project.

- Base URL: https://api.aicu.ai/v1
  OpenAI-compatible. With the OpenAI SDK, just swap base_url and api_key
- Auth:
  Authorization: Bearer $AICU_API_KEY
  Set the AICU_API_KEY environment variable
  Get a key at https://api.aicu.ai/dashboard/keys
  Keys have scopes (llm / tts / images) — include
  tts/images when issuing if you need speech or image generation
- Chat:
  POST /v1/chat/completions
  List models and prices with GET /v1/chat/models
- Text to speech:
  POST /v1/audio/speech
  body {"input": your text, "voice": "nao"} → returns an MP3
  eleven_multilingual_v2 (default) takes up to 10,000 chars per request
  The response is streamed, so writing the response body
  straight to a file yields a complete MP3
- Image generation:
  POST /v1/images/generations
  model "gpt-image-2"
  Supports character-consistent generation from reference images
- Transcription:
  POST /v1/audio/transcriptions
  multipart with file and model=whisper-large-v3-turbo
- Errors are never billed
  Re-running the same input is free whenever the cache can serve it
- Full documentation:
  https://api.aicu.ai/docs

First check that AICU_API_KEY is set, then verify the setup by
running exactly this and confirming the MP3 saves correctly,
before moving on to the real implementation (this exact sample
is cached server-side and returns free and fast):

curl https://api.aicu.ai/v1/audio/speech \
  -H "Authorization: Bearer $AICU_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"input": "Hello! Welcome to AICU API.", "voice": "nao"}' \
  --output hello-aicu.mp3

A 403 with required_scope means the key lacks that scope —
reissue it with the scope at https://api.aicu.ai/dashboard/keys.