Free
Sign up to try the agent. No card needed.
- 2 prompts per day
- Text + voice + visual responses
- Watermark on rendered clips
- Community support
DonkeyClip turns LLM answers into live, narrated, animated explanations on the web.
Press Enter to send · Shift+Enter for a new line
These are real prompts an agent can already answer. Pick one and we'll drop it in the prompt box and play it for you.
Have something different in mind?
Write your own promptEvery response is composed of three streams that run together, so the answer can be read, heard, and watched at the same time, like a tiny narrated explainer.
A clear, conversational transcript that includes the agent's reasoning and the tool calls it made along the way.
Synthesized voice (WebSpeech, or any TTS provider) that explains each step exactly as the canvas changes.
A deterministic timeline of shapes, motion, and color. Replayable, scrubbable, and exportable. It isn't a flat video file.
Drag in the tools you want: diagrams, voices, quiz steps, branches. Give it a persona. Ship it. Tutors, accountants, fitness coaches, data explainers, all running on the same engine.
2 free prompts per day, no credit card. Bring your own LLM key on Pro for zero token overage, or top up a prepaid wallet and we charge tokens at provider cost.
Sign up to try the agent. No card needed.
No daily limit. Pay-as-you-go via prepaid wallet.
Team seats, SLA, dedicated infra.
Every DonkeyClip agent runs on MotorCortex, our own animation engine. We made it because nothing else could play an LLM's answer the way it needed to.
The model never runs code in your viewer's browser. It only describes what should happen on screen, and MotorCortex draws it. Anything outside that vocabulary has nowhere to execute, so a bad prompt can't turn into a bad payload.
Every agent gets its own URL. Drop it in a lesson plan, a Slack thread, or a Discord with ten thousand people. They open the link and the live agent runs in their browser. No install, no account.
Other AI video tools make you wait while something renders in the background. MotorCortex doesn't render anything. Each answer plays the second it's ready, and your viewer can pause it, scrub it, tweak a parameter, or replay it without leaving the page.