Ojin
Build conversational AI agents with lifelike visual avatars
About Ojin
Ojin is a conversational AI agent platform that offers a complete solution with lifelike visual avatars: voice input, voice output, and lip movements and expressions synced to the audio. It provides two modes: Ojin Agent, where Ojin handles speech recognition, large language model, speech synthesis, and avatars end to end, so you only need to configure the persona; and Third-Party Agent, which lets you bring your own voice-to-voice provider (such as Hume, ElevenLabs, or Ultravox), with Ojin handling the avatar. The platform offers a REST API for managing model configurations, assets, and Human Agents, as well as a WebSocket API for real-time media streaming.
Best for
Best use cases
Customer support: deploy human-like agents to provide personalized 24/7 support
Sales: use conversational avatars to greet, qualify, and convert leads
Server-to-server production deployments that need real-time streaming video and audio frames
Features
Lifelike visual avatars
The flagship face model delivers rich expressions and natural movements (including hands), going beyond simple lip sync. A lower-cost real-time lip-sync model is also available, which turns a single reference image into a natural animated avatar with audio-synced lip movement and expressions, with latency under 200ms and up to 720p resolution.
Two build modes
Ojin Agent has Ojin handle speech recognition, large language model, speech synthesis, and avatars end to end. Third-Party Agent lets you bring your own voice-to-voice providers such as Hume, ElevenLabs, or Ultravox, with Ojin adding the avatar.
REST and real-time APIs
The REST API manages persistent resources such as model configurations, assets, and Human Agents. The real-time model API streams media over WebSocket, so you send voice audio and receive synced video and audio frames in return.
Automatic scaling
Infrastructure handles scaling automatically, so users never see service unavailability.
Pros
- Offers two modes: the all-in-one Ojin Agent and the Third-Party Agent that lets you bring your own voice provider
- Infrastructure auto-scales so users never experience service unavailability
- Per-minute billing with no long-term commitment, plus a $10 free credit
- Complies with the EU AI Act, GDPR, and Saudi Arabia's PDPL
Cons
- Real-time WebSocket API is designed for server-to-server use and must be connected from a backend server rather than the frontend
- Real-time API rate limit is 6 requests per second, with a maximum message size of 512 KB
- Video output streams in real time at 25 fps
Pricing Plans
Creator
- First agent
- 10,800 credits per year
- Most popular
Studio
- Production products
- 132,000 credits per year
Growth
- Scaling and high usage
- 408,000 credits per year
* Pricing information is for reference only. Please refer to the latest announcement on Ojin's official website.