60+ Tokens/Sec Live Streaming AI Gateway & Agent Control Plane

HIVV.org operates a distributed AI network where thousands of computers pool their resources to provide free AI access without requiring a signup or data collection. The platform streams language model responses at high token rates, combining on-premise hardware inference with multi-provider cloud failover. It offers an OpenAI-compatible API endpoint and supports agent governance and priority routing for users who want to integrate AI into their workflows.

The service is completely free with no signup required, and it explicitly states that user data stays private by routing messages peer-to-peer rather than through a central server. Users can select from various AI models including on-premise options like Gemma-4 and llama.cpp, as well as hosted models from providers such as Anthropic, OpenAI, Google, and Groq. The platform includes features for agent governance, priority routing, and real-time streaming of token output to ensure fast and private interactions.

Getting started requires no installation or registration; users can simply copy a provided command and run it in their terminal to talk directly to the distributed AI network. The system uses an OpenAI-compatible API endpoint at https://hivv.org/v1/chat/completions where authentication is handled via an API key that can be any string for consumers. Hosts can join as providers by donating a key and running a node to serve an LLM to the global network.

The platform supports multiple routing modes including balanced, quality, fast, cheap, private, and trusted to let users choose how their requests are handled. Users can self-host an endpoint by posting to https://hivv.org/api/keys/donate with a custom provider and API key pointing to their own IP address. The service also provides a CLI guide and OpenAPI documentation for developers who need to integrate the gateway into their applications.

The service claims a throughput of 60+ tokens per second and operates on donated hardware across a global network of computers. It runs an AI control plane that allows for agent governance and priority routing, ensuring that critical tasks can be prioritized over standard requests. The platform also includes a health check endpoint that returns uptime and node status information to help users monitor the network.

The platform supports voice chat capabilities and provides a live status page for users to check the current state of the network. It also includes a policy endpoint at https://hivv.org/api/policy and a capabilities endpoint at https://hivv.org/api/capabilities for developers to understand the available features. The service is designed to be a unified AI gateway and sovereign agent governance plane control center.

The service is described as a global network that runs on donated hardware, making it accessible to anyone with an internet connection. It explicitly states that there is no signup required and that the system is 100% free for all users. The platform aims to provide free AI for everyone by pooling resources from thousands of computers worldwide.

The platform uses a protocol compatible with OpenAI Chat Completions (v1/chat/completions) and provides a self-hosted endpoint option for users who want to run their own instances. It also includes a manifest file at https://hivv.org/ai.txt and an OpenAPI specification at https://hivv.org/openapi.yaml for developers. The service allows users to discover and connect by checking health status and listing available models.

The service is designed to be privacy-focused, with messages going straight to the node without passing through a central server. It supports streaming of token output in real-time, allowing users to see the AI response as it is generated. The platform also includes a CLI guide and a quickstart guide for users who want to begin using the service immediately.

The service includes a donate endpoint at https://hivv.org/api/keys/donate where users can post their provider API keys to join the network as a provider. It also supports custom endpoints for users who want to self-host their own LLMs. The platform is built to be scalable and can handle multiple providers simultaneously.

The platform provides a health check endpoint that returns uptime and node status information, allowing users to monitor the network's performance. It also includes a live status page and a policy endpoint for users to understand the terms of service. The service is designed to be resilient and can handle failures by switching to cloud failover options.

The service supports multiple operating systems including macOS, Linux, WSL, and Windows PowerShell, making it accessible to a wide range of users. It also includes a CLI guide and a quickstart guide for users who want to begin using the service immediately. The platform is designed to be easy to install and use with minimal dependencies.

The official website of 60+ Tokens/Sec Live Streaming AI Gateway & Agent Control Plane is hivv.org.