Pagram is a decentralized protocol designed to serve as an AI backbone for the livestreaming industry. The project aims to provide content creators with a suite of AI-powered tools to enhance video quality, automate production, and create interactive experiences for their audiences. The protocol's infrastructure is built on a decentralized network of compute providers, with the stated goal of offering a censorship-resistant and cost-effective alternative to centralized AI cloud services.^[1]

Overview

Pagram was created to address the growing reliance of live streamers and content creators on centralized platforms and proprietary AI tools. The project's vision is to democratize access to advanced AI capabilities by building a decentralized network where compute resources are crowdsourced and made available to creators. By leveraging a distributed network, Pagram aims to provide the necessary processing power for real-time AI video manipulation without depending on single corporate entities.^[1]​

The protocol is being developed as an open-source framework, allowing developers to build new AI-driven applications or integrate Pagram's functionalities into existing streaming software. The core of the project is to provide a foundational layer for AI-enhanced livestreaming, fostering an ecosystem of tools that prioritize creator control and innovation. The platform is designed to equip streamers with production-grade AI features that were previously only accessible to large-scale broadcasting studios.^[1]​

Products

Pagram provides a set of tools and infrastructure for AI-assisted livestreaming. It combines real-time media processing, automation features, and distributed compute into a single system.

Pagram AI Plugin

The Pagram AI Plugin is designed to be the primary point of interaction for content creators. It integrates with existing streaming software (such as OBS or Streamlabs) to provide a user-friendly interface for controlling AI-powered features without requiring technical expertise.^[1]​

Real-Time Video Enhancement

The plugin offers tools to improve the visual quality of a live feed. This includes features such as AI-powered upscaling, noise reduction, color correction, and lighting adjustments, which are processed in real-time to enhance the viewer's experience.^[1]​

AI-Generated Overlays

Creators can use the plugin to apply dynamic and context-aware graphical overlays to their streams. These can include automatically generated subtitles, real-time data visualizations, or branding elements that react to in-stream events.^[1]​

Automated Production Tools

The plugin includes features designed to automate complex production tasks. This can involve automatic scene switching based on the on-screen content, intelligent camera angle selection in a multi-camera setup, or the automated creation of highlight clips from the live broadcast.^[1]​

Interactive AI Functions

This set of features is aimed at increasing audience engagement. Examples include AI-powered chat moderators, real-time sentiment analysis of viewer comments, and interactive polls or Q&A sessions managed by an AI assistant.^[1]​

Pagram AI Engine

The Pagram AI Engine is the core technological component responsible for executing the AI models that power the plugin's features. It is engineered to handle the intensive demands of real-time video processing with minimal latency. The engine is designed to be portable and can be run on the distributed nodes of the Pagram Network. Optimization techniques are employed to ensure efficient use of GPU resources, allowing the engine to perform complex tasks like object detection, video enhancement, and natural language processing on a live video stream without significant delay.^[1]​

Pagram Network

Pagram's infrastructure is built upon the Pagram Network, a decentralized marketplace for AI computing power. The network connects individuals and data centers willing to rent out their GPU capacity (node operators) with content creators who need to run AI workloads for their livestreams. The benefits of this model include increased censorship resistance, as there is no single point of failure or control, a potential reduction in costs compared to traditional cloud providers, and greater access to high-performance computing for individual creators.^[1]​

Features

The features of the Pagram protocol are delivered through the AI plugin and are powered by its underlying AI Engine. These features are organized into several categories designed to enhance different aspects of a livestream production. ^[2]

Real-Time Video Enhancement

The protocol is designed to improve the visual quality of a livestream in real time. This includes features for automatic color correction to balance and enhance the video's color profile, as well as lighting adjustments to compensate for suboptimal lighting conditions. It also aims to offer frame upscaling to increase video resolution and a function for removing visual noise and compression artifacts, resulting in a cleaner video feed. ^{[1] [2]}

AI-Generated Overlays & Scenes

Pagram intends to provide tools for creating dynamic visual elements. The system is designed to facilitate dynamic layout switching, allowing for seamless transitions between different on-screen arrangements. It also plans to feature real-time subtitle generation, which automatically transcribes spoken words into text on the screen. Another planned feature is the ability to generate animated overlays on-the-fly, providing creators with custom graphics without the need for pre-made assets. ^[2]

Automated Production Tools

A significant part of the protocol's functionality is focused on automating production tasks. One of the key planned features is intelligent camera tracking, which can automatically follow a subject's movement. The system also aims to generate automatic stream highlights by identifying key moments in a broadcast. An "AI director mode" is another intended feature, designed to automate scene switching by analyzing the content and context of the stream to make intelligent decisions about what to show on screen. ^{[1] [2]}

Interactive AI Functions

Ecosystem

The Pagram ecosystem consists of four main components that work in tandem to deliver decentralized AI processing for live video streams.^[1]

• Pagram AI Plugin: A user-facing software component that integrates directly with popular streaming applications. It serves as the interface for creators to access and control the various AI features.
• Pagram AI Engine: The core software that runs the AI models for video processing. It is optimized for low-latency performance to ensure that enhancements and effects are applied in real-time during a live broadcast.
• Pagram Network: A decentralized network of nodes that provide the computational power (primarily GPU resources) required to run the Pagram AI Engine.
• PGRM Token: The native utility and governance token of the Pagram network, used for incentivizing node operators, paying for services, and participating in governance.

Use Cases

Pagram enables creators, developers, and platforms to enhance, automate, and scale livestreaming through AI-powered tools and decentralized infrastructure. Below are the core ways the in ecosystem can be used:

• Real-time video enhancement — Improve stream quality with auto color correction, lighting adjustments, upscaling, and noise reduction
• Automated livestream production — Reduce manual effort with AI-driven camera tracking, scene switching, and highlight generation
• Dynamic overlays & scenes — Create and switch layouts, subtitles, and animated overlays in real time
• Audience engagement tools — Enable chat moderation, AI-powered Q&A, and personalized interaction prompts
• AI-powered media processing — Enhance video, audio, and interactions using vision, speech, and NLP models
• Decentralized compute access — Run AI workloads on distributed nodes for lower cost and improved scalability
• AI asset marketplace — Buy, sell, or rent overlays, models, and automation tools within the ecosystem
• Creator-to-creator economy — Support direct transactions and collaboration between users
• Node participation & rewards — Earn tokens by contributing compute resources to the network
• Secure & private streaming — Protect data with encryption, sandboxing, and zero-knowledge verification
• Flexible platform integration — Integrate AI features into existing streaming tools via modular plugin

Architecture

Pagram is designed as a modular system composed of layers for user interaction, media processing, and distributed computation. This structure allows components to function independently while remaining integrated, supporting real-time livestream workflows across different environments. ^{[1] [2]}

The client layer consists of a plugin that integrates with livestreaming software. It handles media input, applies selected features, and manages output during live sessions. It also supports user configuration and interaction, with some processing tasks executed locally to reduce latency where possible.

The AI processing layer provides the models used for media enhancement and automation. It includes video functions such as upscaling and frame adjustment, audio features like noise suppression and speech enhancement, and additional capabilities in natural language processing and computer vision. Models are optimized through techniques such as model distillation and adaptive selection to maintain performance across varying system conditions. ^{[1] [2]}

The compute layer extends processing through a distributed network of nodes. These nodes provide computational resources for executing workloads beyond the local environment. A coordinator manages task allocation based on availability and latency, while isolated execution environments help maintain system integrity. ^[1]

Data flows from capture at the client layer through processing and back for real-time output, without requiring persistent storage of livestream content. Security measures such as encrypted communication and sandboxing are applied throughout. ^[1]

Overall, the architecture separates responsibilities across layers to support flexibility, scalability, and consistent operation across different use cases. ^[1]

Tokenomics

The Pagram ecosystem is powered by its native utility and governance token, known by the ticker $PGRM .^[1]​

Token Allocation

Total Supply: 1,000,000,000 $PGRM

• Private Sale: 5%
• Public sale: 10%
• Team: 15%
• Liquidity: 7%
• Marketing: 15%
• Advisory: 3%
• Staking: 15%
• Treasury: 10%
• Community: 20% ^[1]

Utilities

The $PGRM token is designed with several key functions to facilitate the operation of the decentralized AI network.^[1]​

• Compute Payments: The primary utility of the token is as a medium of exchange for paying node operators for the GPU resources used to process AI tasks.
• Staking: Node operators are required to stake $PGRM tokens as a form of security deposit to ensure their honest participation in the network. In return for their services, they earn staking rewards and fees.
• Governance: Token holders can participate in the Pagram DAO, allowing them to propose and vote on network upgrades, changes to the fee structure, and the overall strategic direction of the project.
• Access to Features: The token may be used to unlock premium AI models, advanced features, or higher processing priority within the Pagram AI Plugin.

Governance

Governance of the Pagram protocol is conducted through the Pagram DAO. Holders of the $PGRM token can participate in the decision-making process, ensuring that the development of the network is aligned with the interests of its community of creators, developers, and node operators. This model is intended to decentralize control over time, moving authority from the initial development team to its stakeholders.^[1]​