Pre-proposal: Creative Industries SPE

origin · June 6, 2025, 4:32pm

Creative Industries SPE Proposal

Request: 12,500 LPT • Duration: 4 months ending ~ October 25, 2025

Prepared by: Sean Bradford

Contact: .originstory / sean@originstories.live

The goal of this SPE is to enhance network capabilities for creative industry use and to demonstrate impactful pilots with key partners to drive demand and grow the Livepeer ecosystem.

Previous Work & Impact

Since 2023 we have been working to build novel
experiences in the creative sector that leverage the Livepeer Network.

Our E² pilot successfully demonstrated Livepeer’s potential in creative applications and clear opportunities for growth.

Educational Content: Created comprehensive video content covering the AI conversation to help spotlight Livepeer to new audiences. Over 20k views and more than 600k impressions since the content launch. The goal of this was to create evergreen content and content frameworks that can continue to be used as the industry evolves.
Public Experiences and Community Engagement: Generated significant interest in creative AI applications through workshops and demonstrations across the creative sector. 12+ hours of generation across Livepeer products.
Technical Validation: Proved feasibility of creative AI workflows on Livepeer infrastructure and a clear understanding of where improvements are needed.
Partnership Development: Established relationships with key creative industry professionals and educators

Building on these proven foundations, we aim to bring new capabilities to the network and scale proven concepts in high-demand market segments.

Abstract

Livepeer has a unique advantage due to the presence of solid infrastructure for both AI and video jobs.

However we believe there are still additional capabilities that are needed in order for the network to increase usage in key markets such as the $3 Trillion entertainment industry.

We request 12,500 LPT to scale real-time AI and batch AI video pipelines on Livepeer tailored for creative use. In a focused 4-month sprint, we will create a Bring Your Own Container (BYOC) endpoint that brings state of the art models to the network and run three high-impact, open-source proof of concepts targeting massive-scale markets:

Conversational AI for Self-Taping (40,000+ daily self-tape auditions market, testing with established Film,Theatre, and TV partners)
Broadcast Events Tool (testing with productions reaching 1.17M+ viewers with access to a network that has 30M+ viewers a year)
Music Visualisation B2B2C Showcase (Focused showcase to drive demand and engagement at Amsterdam Dance Event, a major industry conference). This is a free experience also allowing to help shift public perception.

We expect to onboard 25 + high-quality creative companies and projects, with a minimum usage of 8,000 minutes of usage during pilots with a clear roadmap to 50,000+ minutes over the next 12 months driven by market-scale adoption.

Our BYOC endpoint will contain at least 6 state of the art models across conversational AI, translation, and image/video generation.

These efforts will drive immediate network usage with a focus on scalable applications. We will work to onboard potential long-term creative partners, and validate Livepeer as the premier decentralized platform for AI-augmented media.

Our efforts will also make it easier to onboard and serve incoming/existing creative partners by offering a modular environment that can be used across a variety of use cases with proven demand scalability.

Mission

Building on proven results from our E² pilot, we aim to:

Scale proven creative applications to mass-market segments with demonstrated demand
Prove Livepeer can handle enterprise-level, scalable AI workloads through BYOC & Comfystream
Create clear pathways for developers and production companies to integrate Livepeer into high-volume workflows
Help establish Livepeer as the go-to infrastructure for AI-augmented media production at scale

Rationale & Market Scale

Large Addressable Markets: Rather than boutique applications, we’re targeting markets with proven scale:

Self-Taping Market: 40,000+ daily audition self-tapes requiring AI coaching and feedback. This is an example use case for conversational AI.
Live Broadcast Production: Multi-million viewer events needing real-time AI enhancement
B2B Creative Event Technology: Developer-focused solutions for enterprise creative workflows in music and live events.

Technical Barriers: Creative professionals face significant technical barriers when attempting to integrate AI models and video processing into their workflows. They need pre-configured, accessible solutions rather than raw technical capabilities.
Model Relevance: Out-of-the-box AI models often don’t address the specific needs of creative applications, requiring customization or fine-tuning for optimal results.
Demonstration to Developers: Creative professionals and their technical teams respond best to concrete demonstrations of capabilities rather than technical specifications, especially at B2B conferences.

Technical Approach

Our core technical contribution will be developing and implementing a Bring Your Own Container (BYOC) endpoint specifically designed for creative industry applications. This approach offers several advantages:

Flexibility: The BYOC approach allows us to deploy and test specialized models without waiting for them to be natively integrated into the Livepeer network.
Relevance: We can select and optimize models specifically for creative industry use cases, ensuring they address real needs rather than generic capabilities.
Accessibility: By creating standardized endpoints, we lower the technical barriers for creative professionals to access and utilize these models.
Scalability: The container-based approach allows for easy scaling as demand increases, providing a path from pilot to production.

We will work with Brad and also reach out to orchestrators to run our BYOC.

Comfystream, GWID & Daydream

We will use Comfystream for real-time visual workflows, and any models that are not possible with BYOC.

As the Gateway Wizard progresses we will also evaluate and document these experiences for other projects to be able to use in the future.

Towards the end of this sprint we will also evaluate whether or not we should make our own gateway.

Key Features

Conversational AI: real-time dialogue, low-latency audio/video pipelines, creative-context optimization
Batch Model Support: high-quality image/video generation workflows for style transfer, avatar creation, scene extensions, etc.
Real-time Processing: Comfystream-enabled pipelines for live events, optimized for performance and latency. We will work with the Broadcast production partner to test translation, and other key optimization features.

Technical Implementation

The BYOC endpoint will include:

Conversational AI Pipeline

Implementation of state-of-the-art conversational model optimized for creative contexts
Integration with audio/video streams for real-time interaction
API endpoints for easy integration with existing creative tools
Performance optimizations for low-latency responses

Batch Image and Video Generation

Integration of leading image and video generation models to improve production quality
Optimization for creative industry use cases (e.g., style transfer, scene extension, visual effects)
Efficient batch processing capabilities

Real-time Processing Capabilities

Utilization of Comfystream for real-time AI applications and Livepeer, Inc applications when appropriate
Optimization for creative contexts like live performances and broadcasts

Pilot Projects

Pilot	Description	Target Users	Market Scale	Minutes
Conversational AI for Self-Taping	Interactive AI coaching assistant for audition preparation and execution. Working with Broadway, LA, and other relevant professional networks to test demand with actors who create 40,000+ daily self-tapes. Studios and rehearsal facilities represent additional massive scale opportunity.	Actors, casting directors, acting coaches, studios	40,000+ daily self-tapes, unlimited studio/rehearsal potential	2,000
Broadcast Events AI Enhancement	Real-time AI enhancement for live productions tested with beauty pageant reaching 1.17M+ viewers. Production company produces massive live shows, providing pathway to enterprise-scale adoption.	Broadcast producers, streaming platforms, live event companies	Multi-million viewer events, enterprise production companies	2,000
Music Visualisation B2B Showcase	Developer-focused demonstration of BYOC capabilities at Amsterdam Dance Event (major B2B music & tech conference). Designed to show technical teams how to integrate Livepeer, leading to their own implementations.	Developers, music-tech companies, creative agencies, venues	Venue, Organizers and B2B developer adoption leading to their own scaled implementations	3,000

Clear Usage Scaling Pathway

Phase 1 (8,000 minutes): Proof of concept with key partners

Phase 2 (6-month projection - 25,000 minutes):

Self-taping scales to 1,000 regular users
Broadcast production company adopts for 5 major events
ADE showcase leads to 10 developer teams building their own applications or partner implementations.

Phase 3 (12-month projection - 50,000+ minutes):

Self-taping becomes standard tool across acting community (5,000+ regular users)
Broadcast partner scales to monthly enhanced live events
B2B showcase generates 10+ independent developer implementations
BYOC infrastructure enables other creative applications to launch easily

Timeline & Milestones

June/July

Design, build, and test BYOC endpoint
Select models, template orchestration, container validation
Finish Broadcast Events Demo with industry experts.

August

Launch Conversational AI pilot—deploy & gather usage data
Test Broadcast Events demo with the network, production teams, and key stakeholders

September – October

Execute BYOC and Network showcase during ADE
Host Creative Industries workshop leading up to ADE
Use new models and pipelines in 3 month visual live stream with a YouTube channel with 200k+ subscribers (TSG Music)
Collect feedback metrics, document usage and track network fees

By Oct 25

Finalize delivery of all three pilots
Publish reports and code repositories
Share full transparency report with Livepeer community

Budget Breakdown (12,500 LPT)

Category	LPT	Description
Dev & BYOC Build	4,000	Endpoint, container design, API, stability testing
Network Credits	500	Inference to help test and execute the pilots laid out in this proposal
Orchestrator Support	500	Incentive for Orchestrators to run our BYOC
Pilots	4,250	Conversational (1.5k), Broadcast (1k), ADE (1.75k)
Partnerships	750	Focused partner workshop to showcase BYOC and pilots with Swedish innovation companies and stakeholders
Operations/Admin	2,500	Project management, legal, reporting

Success Metrics

Immediate Impact (4 months):

8,000 Livepeer minutes across pilots
3 open-source demos with code/docs
25+ creative partners onboarded
3 published case studies

Scaling Indicators:

Theatre, Film, TV network usage and public engagement around self-tape solution.
Broadcast production company commitment to ongoing usage
Number of developers/companies requesting BYOC access post-ADE
Network fee generation trajectory

12-Month Projections:

50,000+ minutes of validated network usage
Self-sustaining adoption in self-taping market segment
Enterprise broadcast partnerships established
Developer ecosystem actively building on BYOC infrastructure

Team & Governance

Core Team

Sean Bradford — Project Lead
Entertainment industry veteran across music, theatre, and film. Co-created the first blockchain-enabled generative AI tool, which saw 70k+ generations and 8,500+ users during its prototype phase. B.S. Northwestern University (7.2% acceptance rate), 2× NY Times feature, 20M+ Spotify streams. Member of Stable Founders and AI Sweden. Extensive educational background, including co-founding an international arts education non-profit.
Joseph Wan — UX/UI & Integration
Industrial product, data-driven, and UX designer with a passion for harnessing AI-enabled capabilities to redefine human-object interaction. Master’s in Data-Driven Design from the University of Applied Sciences.
Portfolio → josephw.space
Sneha Gupta — Full-Stack Developer
Electronics and Communication Engineer, proficient in deep learning and computer vision, and full-stack web development.
Additional Backend Developer (To be announced)
Additional AI/ML Support (To be announced)

Partners

Adam Kantor
Tony, Grammy, and Emmy Award-winning actor, producer, and coach. Helping oversee design and early testing of the conversational AI use case.
XIXI
A global production company involved in major events from music awards shows to the G20 Summit, reaching over 30 million viewers annually. Team experience includes Netflix and top-tier South African broadcast productions.
Thijs Verhulst — Angle of Impact / Music Tech Netherlands
Former Warner Bros. Benelux executive. Supporting ecosystem growth and facilitating key implementation conversations.
Music Tech Netherlands → musictechnetherlands.com
Nordic Music Tech / Stockholm Ecosystem
Led by CEO Jenny Björk, fostering innovation in music and tech across the Nordics.
TSG Music
Led by Almer and team. Supporting music visual livestreams ahead of ADE and partnering with collectives like The District to provide access to top and emerging tracks in dance music.

Advisors

Andrew Dubber - With a deep technical network and experience in creative innovation through MTF Labs and his Industry Commons initiative, Andrew is able to advise on opportunities, technical needs, and market readiness.

Governance

Funds will be disbursed by the core SPE team
Monthly progress reports will be delivered starting end of July
Monthly presentations in Livepeer community of developments and usage.
Community updates will be posted via our core-contributor channel in the Livepeer forum

Community Collaboration

We are committed to working closely with the broader Livepeer ecosystem including coordinating with the AI SPE where needed, The Foundation and Marketing Teams around broader coordination.

We are also open to work with our creative SPE’s and community members to help maximize the impact of our efforts.

Sustainability Plan

BYOC container and docs are open-source
Templates & guides are created for easier use by other creative teams
Lay foundation for future Livepeer network usage, with potential of creating a Creative Industries gateway

Summary

This 4-month sprint is designed to deliver measurable network engagement, validated creative use cases, and strong open-source infrastructure. We believe that with this team and timeline we can help unlock new sustainable demand opportunities for the Livepeer network.

origin · June 18, 2025, 10:07pm

This is the final version of the proposal that will go into the explorer. It includes:

A significantly reduced ask. The current price at posting is lower than the 30-day SMA so we will keep it to 12.5k LPT
A streamlined focused on creating a public good for the network via a creative industries playground
A clearer roadmap of exactly how we will amplify the new capabilities and drive usage to the network

Preface

The goal of this SPE is to enhance network capabilities for creative industry use and to demonstrate impactful proof of concepts that will drive demand and grow the Livepeer ecosystem.

To do this we will build an open-source playground with Livepeer pipelines that make it easy for others to build at the application layer. There will also be a set of 6-8 template workflows that can be used across a variety of use cases.

We will leverage the existing capabilities of the AI SPE with batch pipelines and continue to leverage new primitives like Comfystream, BYOC, and GWID to accomplish the deliverables laid out in this proposal.

Abstract

We request 12,500 LPT to build a playground environment for the Livepeer Network, specifically tailored for creative brands and businesses, and to showcase it across three proof of concepts. This covers engineering, creative development, pilot execution, and documentation.

Livepeer has a unique advantage due to the presence of solid infrastructure for both AI and video jobs.
Our pilot around Education and Experience earlier this year successfully demonstrated Livepeer’s potential in creative applications and clear opportunities for growth.

However, we believe there are still additional capabilities that are needed in order for the network to increase usage in key markets such as the $3 Trillion entertainment industry.

In a focused 4-month sprint, we will create a PWA that brings state of the art models to the network and run three high-impact, open-source proof of concepts targeting massive-scale markets:

Conversational AI for Auditions and Rehearsal (40,000+ daily self-tape auditions market, testing with established Film, Theatre, and TV partners)
Music Visualisation and Optimisation for Labels, Brands, and Organizers. Amplified via B2B2C Showcase (Focused showcase to drive demand and engagement at Amsterdam Dance Event, a major industry conference). This is a free experience also allowing to help shift public perception.
Broadcast Events Tool (testing with productions reaching 1.17M+ viewers with access to a network that has 30M+ viewers a year)

We expect to onboard 25 + high-quality creative companies and projects, with a minimum usage of 8,000 minutes of usage during three proof of concepts. We also anticipate 50k-100k+ minutes of usage over the next 12 months driven by market-scale adoption.

The playground will contain at least 6 templated workflows that incorporate state of the art models across conversational AI, translation, and image/video generation workflows.

These efforts will drive immediate network usage through the three proof of concepts laid out above.

Our efforts will also make it easier to onboard and serve incoming/existing creative partners in the Livepeer ecosystem. by offering a modular environment that can be used across a variety of use cases.

Mission

Building on results and insights from our E² pilot, we aim to:

Scale-proven creative applications to market segments with demonstrated demand
Prove Livepeer can handle enterprise-level, high production quality, and scalable AI workloads
Create clear pathways and documentation for creative groups and production companies to integrate Livepeer into high-volume workflows
Help establish Livepeer as the go-to infrastructure for decentralized AI-augmented media production at scale

Rationale & Market Scale

Large Addressable Markets: Rather than boutique applications, we’re creating a custom environment that can be used across the creative industries:

Self-Taping Market: 40,000+ daily audition self-tapes requiring a real-time scene partner. This is an example use case for conversational AI.
Live Broadcast Production: Multi-million viewer events looking for ways to enhance production with AI and extend creativity to audiences.
B2B Creative Event Technology: User-friendly solution solutions for independent and enterprise creative workflows in music and live events.

Technical Barriers: Creative professionals face significant technical barriers when attempting to integrate AI models and video processing into their workflows. They need pre-configured, accessible solutions rather than raw technical capabilities.

Model Relevance: Out-of-the-box AI models often don’t address the specific needs of creative applications, requiring customization or fine-tuning for optimal results.

Demonstration to Industry: Creative professionals and their technical teams respond best to concrete demonstrations of capabilities rather than technical specifications, especially at B2B conferences.

Technical Approach

Our core technical contribution will be developing and implementing an open-source Livepeer playground specifically designed for creative industry applications. The final version will have the potential to be its own gateway.

This approach offers several advantages including:

Relevance: We can select and optimize models specifically for creative industry use cases, ensuring they address real needs rather than generic capabilities.
Accessibility: By creating standardized endpoints, we lower the technical barriers for creative professionals to access and utilize these models.

We will work with the AI SPE to make sure everything is properly coordinated and have discussed directly with two orchestrators to help run models.

Comfystream, GWID & Daydream

We will use Comfystream for real-time visual workflows, and any configurations that are not possible through existing batch generation or video pipelines.

As the GWID project progresses, we will also evaluate if we can use this to offer a fiat on-ramp for usage. We will document our results and experiences for other projects to be able to use in the future.

Key Features

Conversational AI: real-time dialogue and voice cloning and synthesis.
Batch Model Support: high-quality image/video generation workflows for style transfer, avatar creation, scene customization and other use cases.
Real-time Processing: Comfystream-enabled pipelines for live events, optimized for high-quality results personalized experiences.

Technical Implementation

The playground will include:

Conversational AI Pipeline

Implementation of a state-of-the-art conversational model optimized for creative industry needs.
API endpoints for easy integration with existing creative tools

Batch Image and Video Generation

Integration of leading image and video generation models to improve production quality. These steps will be coordinated with the AI SPE.
Optimization for creative industry use cases (e.g., style transfer, scene extension, visual effects)

Real-time Processing Capabilities

Utilization of Comfystream for real-time AI applications and INC applications when appropriate
Optimization for creative contexts like live performances, broadcast events, and large-scale experiences

Playground Proof of Concepts

POC	Description	Target Users	Market Scale	Minutes
Conversational AI for Self-Taping	Interactive AI coaching assistant for audition preparation and execution. Working with Broadway, LA, and other relevant professional networks to test demand with actors who create 40,000+ daily self-tapes. Studios and rehearsal facilities represent additional massive scale opportunity.	Actors, casting directors, acting coaches, studios	40,000+ daily self-tapes, unlimited studio/rehearsal potential	2,000
Broadcast Events, Live Events AI Enhancement	Template workflows and real-time AI video for live productions and audiences will be tested with large-scale production company XIXI Studios, onboarded during the original pilot for E-squared. The demo will be made specifically for Miss South Africa pageant and presented for use across their various projects which have more than 5 million viewers per year. These same concepts will be tested across a live theatre production at the Edinburgh Fringe Festival.	Broadcast producers, streaming platforms, live event companies	Live productions. Multi-million viewer events, enterprise production companies	3,000
Music Visualisation B2B Showcase	Strategic demonstration of playground capabilities at Amsterdam Dance Event (major B2B music & tech conference). Designed to show technical teams how to integrate Livepeer, leading to their own implementations.	Developers, music-tech companies, creative agencies, venues	Venue, organizers and B2B developer adoption leading to their own scaled implementations	3,000

Outlook for Usage (0-12 months)

Phase 1 (8,000 minutes):

Proof of concept with key partners for:
Self-taping and audition needs: Acting schools, casting agencies, and production teams
Demo and presentations for Broadcast Event and Network Production teams tied to a physical real-world event and pipeline of potential use cases
Amplification of the playground and proof of concept showcase for key music industry professionals

Phase 2 (6-month projection - 25,000 minutes):

Self-taping usage scales due to community adoption, word of mouth and high-caliber industry testimonials
Broadcast production company adopts for 5 major events
The Creative Industries Playground helps other creative applications launch the Livepeer Network
Planned usage for upcoming large-scale experiential project (multi-month)

Phase 3 (12-month projection - 50k-100k+ minutes):

Confirmed experience that will use at least two of the playground workflows - estimated to be used by over 200k people
Self-taping tool used by theatre and film professionals, casting, studios, rehearsal, teams, and more.

Timeline & Milestones