$5 free credits when you sign up Claim now
Wan 2.2 Animate now available Test it!
Video Upscaling models now available Test it!
Z-Anime image model Test it!

OpenClaw Agent Voice, Hearing, Vision & Memory

Give your OpenClaw agent (formerly ClawdBot) every sense it needs — text-to-speech, transcription, image generation, and embeddings. One API key, zero GPU setup. Pay-per-use at up to 20× lower cost than centralized providers.

Get $5 Free Credits Read the Docs

Give Your ClawdBot Every Sense It Needs

One API key unlocks voice, hearing, vision, and memory for your OpenClaw agent. Check the full model list.

  • Text to Speech

    Kokoro TTS

    Agent speaks back in 40+ natural voices via Kokoro TTS.

  • Transcription

    Whisper Large V3

    Understands voice messages and video content via Whisper Large V3.

  • Image Generation

    FLUX, Qwen Image

    Creates visuals from prompts with FLUX and Qwen Image models.

  • Embeddings & Memory

    BGE M3

    Builds long-term memory and knowledge bases via BGE M3.

Up and Running in 3 Steps

From zero to a voice-enabled OpenClaw agent in minutes.

  1. Install OpenClaw

    Follow the setup guide at openclaw.ai. No GPU required on your machine.

  2. Add Your deAPI Key

    Grab your free key and paste it into the OpenClaw config file. One key covers all AI capabilities.

  3. Start Using AI

    Ask your agent to transcribe a voice message, speak a response, or generate an image. It just works.

See OpenClaw + deAPI in Action

A freshly installed OpenClaw agent gains voice, ears, and vision in three config edits — and immediately starts replying with audio, transcribing voice notes, and generating images from prompts.

  • One deAPI key unlocks every capability
  • No local GPU — inference on decentralized infrastructure
  • Works with self-hosted OpenClaw out of the box
  • Works with OpenClaw out of the box
  • No credit card required

Start free, build
something real

$5 credits included. No subscription, no GPU headaches.

Frequently Asked Questions

Everything you need to know

deAPI is a decentralized AI inference API that gives your OpenClaw agent voice, hearing, vision, and memory. One API key unlocks text-to-speech (Kokoro, QwenTTS, Chatterbox), transcription (Whisper Large V3), image generation (FLUX.2 Klein, Qwen Image, Z-Image), video generation (LTX & LTX-2), embeddings (BGE M3) and more — all via simple REST calls.
After creating a free deAPI account and copying your API key, open your OpenClaw configuration file and paste the key in the relevant provider field. Restart OpenClaw and the new capabilities are available immediately.
OpenClaw can access Kokoro TTS (40+ natural voices), Whisper Large V3 (transcription from audio/video), FLUX and Qwen Image Edit (image generation and editing), and BGE M3 (embeddings for long-term memory). More models are added regularly — check the full model list.
deAPI is pay-per-use with no subscriptions. New users receive $5 in free credits — enough for thousands of TTS characters, hours of transcription, or hundreds of generated images. Prices are up to 20× lower than centralized providers like OpenAI.
deAPI runs on decentralized consumer-grade GPUs, delivering open-source models at a fraction of the cost. You get specialized capabilities not available through OpenAI (video generation, background removal, upscaling) and no vendor lock-in.
Follow the setup instructions at the official OpenClaw documentation. Configure your environment variables (including your deAPI key) and you're ready to go. No GPU required on your side — deAPI handles all inference in the cloud.