Pre-proposal - E² - Education & Experience SPE (Phase 2)

[origin]

E² Phase Two: Real-World AI Integration in Creative Sectors

Pre-Proposal

Requested Budget: 28,000 LPT
Prepared by: ORIGIN STØRIES Team
Contact: .originstory / hello@originstories.live

Overview

This six-month extension of the E² – Education & Experience Focused SPE builds on the foundations laid out in the pilot. The goals remain the same, to combine education and experience with the focus of driving innovation and accelerating Livepeer adoption in the creative industries. This phase will specifically focus on high-impact, real-world pilots. By executing AI video workflows in film dubbing, real-time AI visual experiences, broadcast/streaming, and partner workshops, the SPE will expand education, showcase decentralized infrastructure in action, and support long-term ecosystem engagement. You can view our reports from the pilot here.

Motivation

Creative industries are eager to adopt AI tools but lack accessible, context-relevant workflows. Open-source infrastructure like Livepeer can help SMB and creative brands to bridge the gap between potential and practice. This proposal delivers hands-on pilots and educational resources to drive adoption and innovation. There are multiple use cases at different stages of production, providing different levels of value.

Rationale

By partnering with SMB teams and brands in Media and Entertainment, this SPE builds bridges between infrastructure and usage. It allows creative practitioners to test, adapt, and grow their own AI workflows — all while using Livepeer as the backbone for real-time and batch processing. Funds will be distributed on a project basis by the core operations team.

SPE Governance Structure

• Core Team (ORIGIN STØRIES): Responsible for project leadership and implementation, funding distribution, and reporting. Led by Sean Bradford, Sneha Gupta (Full-Stack), and Joseph Wan (UI/UX), plus specialists for AI/ML, Distribution, and regional business environments.

• Key Partners: Leading and contributing to the execution of specific use cases

  • Triangle Drama and Untitled Pictures- Lipsync Dubbing for film (in pre-production, shooting starts 9 June.
  • XIXI Productions - Production company leading Broadcast implementation, initial demo for Miss South Africa, entry point to live production and broadcasts in South Africa, and for global broadcast infrastructure - August
  • Joe Hay, City of Adelaide - Leading presentation and pilot for UNESCO Creative Cities of Music Network.
  • Jenny Bjork Leading strategic educational workshop in Stockholm alongside Nordic Music Tech
  • Thijs Verhulst - Angle of Impact, supporting the ADE Experience.
  • TSG Music - Leading livestreams pre-ADE and partnerships with music collectives such as The District
  • Further Experimentation in the food sector with The Food Office , Olivia Secher, and others

Key Use Cases

• Lipsync-dubbing and translation (hybrid workflow) Doctor Glas (Feature Film, Sweden - Triangle Drama and Untitled Pictures

• Real-time AI, visual and LLM workflows for Broadcast - XIXI Productions - demo target is Miss South Africa (1.17 M Viewers)

*Hands-on experimentation of AI use cases using Livepeer - Workshops with Nordic Music Tech, City of Adelaide, and Partners

• Realtime Interactive Animations in Physical Venues (e.g., restaurants, events) Amsterdam Dance Event + Development with Food Office and Olivia Secher

• Projection-Based Animation via Text Prompts (music experiences) - Livestream series with TSG Music (197K Subscribers)

We will use Comfystream and existing pipelines, while also aiming to work closely with the Cloud SPE and AI SPE to ensure additional pipelines are warm and loaded for each use case.

Timeline & Deliverables

Month	Activities & Milestones
May/June	- Dubbing pipelines added, pre-production setup for feature in Sweden - Demo creation and pitching for Miss South Africa and SABC Broadcasts - Workshop with City of Adelaide + presentation for UNESCO conference - ADE Planning
July	- Bi-weekly livestreams with TSG Music leading up to ADE begin
August	- Public real-time visual testing during two summer events - Focus on customization and easy user interface - AI Reader testing with users in target markets, using Livepeer pipelines
September	- Workshop in partnership with Nordic Music Tech
October	- ADE Experience – extending creativity to audiences via co-creation and real-time AI Visuals - Integration into www.vpm.studio
November	- Reporting, interviews, future strategy planning

Estimated Budget Breakdown (28,000 LPT)

Category	Purpose	Amount (LPT)
Dubbing Workflow (Film)	Voice + lip-sync test with feature film and production partners. Outcome is a template workflow for other projects.	5,000
ADE Experience	Real-time visuals, venue support, documentation.	5,500
Broadcast Pilot (SA)	Demos, TV network testing, Livepeer integration.	2,000
Edu Workshop and Experience - Stockholm	Strategic event introducing Livepeer and Open-Source plus applications and capabilities.	2,000
Real-time AI visuals	Two documented public tests during summer season focused on user experience and engagement.	2,000
Reader Prototype	Further development and pilot testing of tool to relevant users.	1,500
UNESCO Presentation & Adelaide Pilot Preparation	Presentation in June for UNESCO Creative Cities, plus further planning for pilot activation with the city of Adelaide (currently looking at Adelaide) Fringe in 2026.	1,500
Operations & Admin	Team, strategy, legal, compliance, logistics	8,500
Total		28,000

Outcomes

• 6 Livestreams with generative visuals and real-time workflows in partnership with TSG Music and other partners in formation.
• 1 Lip-sync dubbing pilot with feature film partner, Triangle Drama and Untitled Pictures
• 1 Interactive visual experience tied in conjunction with Amsterdam Dance Event
• 1 demo for Miss South Africa broadcast presented with the intention of using in a pilot. This demo will be used to further pitch Broadcast experiences in the South African Broadcast Corporation network as well as abroad (looking next at SVT - Sweden.
• 1 presentation UNESCO creative cities network with the intention of creating a larger-scale pilot activation, in partnership with the city of Adelaide
• 1 strategic education workshop, highlighting Livepeer technology and ecosystem, presented in partnership with Nordic Music Tech
• 2 documented public tests around real-time AI visuals in summer.
• Expanded workshop and development in the food sector around real-time and other valuable use cases, in collaboration with domain-specific design partners.

Community Collaboration & Livepeer Marketing Support

At the end of this phase we would like to have a framework for tracking AI usage, which can also be use to onboard partners. This can help us as we test pilots and work with specific projects to gauge and monitor demand and growth on the network

Over the course of the next six months we would like to collaborate with AI SPE as well as the Cloud SPE in terms of required pipelines and to further explore collaborations around agentic workflows with the Agent SPE. We also think there is opportunity for some collaboration on specific activations with with some of the creative and media focused SPE’s.

In terms of Livepeer, it would be great to collaborate with marketing efforts around specific activations and announcements to further amplify awareness, and showcase the potential and keyplayers in the network.

Transparency & Accountability

• We will create a public discord channel to post updates and communicate with the Livepeer ecosystem.

• We will post monthly updates and issue quarterly reports outlining progress and challenges.

• KPIs: Our main KPIs are focused on pilot completion and demo delivery across each use case. We will also be tracking usage for planned experiences such as the livestreams leading up to ADE, and the experience during ADE in October.

  • Pilot Deliverables Completed
    Target: 6 successful use case demos delivered (one per core pilot)
  • Workshop & Event Reach
    Target: 1,000+ total attendees and participants across workshops, livestreams, and live activations
  • New SMBs or Partners Onboarded to Livepeer
    Target: 10+ SMBs or organizations introduced to and actively testing Livepeer tools
  • AI Video Use and Pipeline Interactions
    Target: 2,000+ minutes of AI video processing and 10,000 LLM/AI pipeline interactions

Conclusion

This proposal leverages Livepeer’s infrastructure and network to deliver practical, high-visibility pilots in dubbing, immersive events, broadcast, and education to drive real-world. By showcasing applied pipelines for each use case, we aim to foster innovation, drive adoption and bring traditional media and entertainment players into the Livepeer network.

[origin]

Adding some extra information based off of feedback.

Here is the pilot proposal from last Dec
Our final pilot reports

Key Terms:

SMB - Small/Medium Business - This can be brands, production companies, creative studios, artists/streamers/contentcreators who are brands.

If there are other key terms that need elaboration, please let me know.

In terms of key results in demand we expect it to be higher than the amount included above coming in closer to 8000 minutes of pipeline usage

We are anticipate this will lead to 50,000+ minutes of AI video processing within 12 months through a combination of real-time experiences, post-production dubbing, and follow-on adoption.

Initial outputs from dubbing workflows, livestream activations (ADE, Pride), broadcast pilots, and tool testing account for 8,000 minutes.

An additional 42,000 minutes is projected from 30+ SMBs and productions adopting templated workflows, in music film, and broadcast, along with educational workshops, strategic partnerships, and co-created campaigns.

[origin]

Ok this will be the final revisions based on additional feedback. Here we are revising the name to CREATIVE INDUSTRIES SPE which aligns more with the Livepeer future vision. We also included a more concrete technical breakdown around what is being built.

Based on another round of feedback we will change the name of this SPE, to Creative Industries SPE.

This SPE is a continuation of the work from the pilot, including further development with initial partners.

Education and Experience are now ingrained as part of our core values, but it is clear we have strong partners, and clear opportunity to bring even more value to the Livepeer network.

Our Goal: To increase usage of the Livepeer Network in the creative industries

We will accomplish this by bringing new SOTA models, leveraging the existing pipelines and developments, and working with our design partners from the pilot phase.

We will bring new pipelines on the network, using and evaluating BYOC (Bring Your Own Container), Comfystream, the Daydream API, and emerging possibilities like GWID. We are also discussing with other groups and are sure we can execute on our proposal. We are also getting advice on these buildouts from senior companies like MDCS.AI to evaluate future needs around data and compute as use cases scale. And through MTF Labs and our advisor Andrew Dubber, we also have access to a large network of technical minds as this project continues to grow.

Our team is also actively working to bring in new technical talent from universities like Delft and KTH,** as well as local talent where initial projects are focused. This will help to put Livepeer in front of new communities.

Deliverables:

The code for our Dubbing workflows will be open-source.

We will work on 6 pilots and onboard at minimum of 50 SMB and Enterprise Leaders to the Network.

We estimate 8,000 minutes in production for this pilot and an increased demand by 50,000 minutes from pilot efforts and future partnerships that occur as a result of these efforts.

The main pilots are being tested with high-visibility real-world use cases with a combined potential demand of more than 10 million minutes a year. You can see more about this in our report on AI in the Creative Industries.

Category	Purpose	Amount (LPT)	% of Budget
Development & Prototyping (Tools)	Reader Prototype, Live AI Visuals, Broadcast Pilot, Dubbing Workflow	10,500	37.5%
Events & Activations	ADE Experience, Edu Workshop Stockholm, UNESCO Presentation & Adelaide Planning	9,000	32.1%
Operations & Admin	Team coordination, legal, logistics, compliance	8,500	30.4%
Total		28,000	100 %

0-3 months

Development and POC of Hybrid Lip-sync Dub

This is for a Swedish feature (Doktor Glas). Note the actors will not be doing voice cloning, but record audio in English to increase production quality and emotional authenticity in there performances.We will also work with a post-production PM with VFX/AI experience.

Will evaluate Comfystream/ComfyUI and BYOC for interacting with he Livepeer Network.

Models we are evaluating include:

numz/wav2lip_studio-0.2 · Hugging Face

openai/whisper-large-v3 · Hugging Face

https://github.com/OpenTalker/SadTalkerhttps://github.com/google-ai-edge/mediapipe

GitHub - EvelynFan/FaceFormer: [CVPR 2022] FaceFormer: Speech-Driven 3D Facial...

The aim is 70% faster than full manual VFX pipeline, 60% lower cost than traditional dubbing + VFX and a scalable approach for future projects

Outcome - Open-sourced code for complete dubbing workflow, attachment of Livepeer Network to a highly anticipated film project, ability to replicate this template for an in-demand market. There is already initial interest from other projects post pilot.

Development and MVP of Demo for interactive Broadcast productions - presenting initial demo for Miss SA (1.17+ M viewers)

This is a combination of existing pipelines on the Livepeer Network as well as ones in development. It also builds upon the templates we’ve developed for co-created experiences. You can view an example repo here (DM me for link)

Outcome - Opportunity to use Livepeer Network in a series of productions throughout the year (30 M+ viewers,

Already in Development - AI Reader

Audition assistant using conversational AI. We will begin testing existing Livepeer models to integrate and creating new workflows with additional models for conversational AI.

Walk-through of Figma

Figma

Prototype in Progress

Testing of UI/UX and customisation in 2 public experiences.

Presentation of work done with Adelaide for UNESCO Creative Cities and at Livepeer Berlin Event

3-6 Months

Workshop for top ecosystem leaders and startups in Stockholm

This will be educational deep dive into AI for Creative industries featuring Livepeer technology. We will encourage usage through credits, showcase the development of the dubbing workflows, and onboard/secure SMB Partnerships

Showcase at Amsterdam Dance Event.

This is a culminate event of the previous work, with the goal of securing partnerships and further increasing demand through real-world examples.

How will measure success?

Through minutes on the network and number of successful production company/venue partnerships.

We have previously built high-quality tools and experiences and thoroughly performed market validation. We have evaluated scope and potential demand, and for some of the test cases have both Figma designs and initial repos started. I

Why is this a public good? We are continuing to educate prospective users of the network through creative use cases that also bring demand. Tech education in creative industries is rare, and a collaboration between the two is at the moment, necessary.

Technical Team.

In terms of engineering, we are planning to add to our existing team, including Sneha Gupta (Full-Stack and Blockchain) and Joseph Wan (UI/UX) and are evaluating candidates from KTH and Delft, facilitated with the help of our Local Leads

Daphanie Oon - Delft University of Technology, NL and

Nicholas - PhD Student a KTH (Stockholm)

In 0-3 months we will focus on development and demos for:

Lip-sync dubbing - This is a hybrid solution which we will either use Comfystream or BYOC in combination with seasoned professionals - Big Indie Swedish Film

• starts shooting 9 June

Broadcast- We are working with producers of Miss SA (1.17 million viewers) on a real-time/batch generation combo app. This show is scheduled for August and demo unlocks access to numerous productions.

We will be experimenting with refinements to the existing real-time video pipelines and applications in two public facing events to confirm market validation. This will be based off of versions of our existing codebase and frameworks.

In 3-6 months

We will onboard more than 50 SMBs to the Livepeer network and have a pathway to repeatable use of 10,000+ minutes a month.

We will conduct a large-scale test and showcase of the advancements in real-time combined with our event template during Amsterdam Dance Event. This is a business conference with strategic goal of increasing future usage. We are working in tandem with Music Tech Organizations and large industry players to maximise impact including partners in our key partner list.

Really excited to bring this foward and help grow the Livepeer Network

[origin]

We are just following up in the forum per Everest
Nodes request and an against vote from @vires-in-numeris

At the top we want to highlight that the subgroup we serve which is media and entertainment is a 3 trillion + established industry which can generate well over 30 million minutes a year in demand when capturing even a small segment of the network
point.

The educational videos have been out for more than a month with 20k+ views on YouTube, Instagram, and Tik Tok combined -

ORIGIN STØRIES

We also issues a full report of how we will drive demand after onboarding and working with industry players and strategising with them the appropriate plan of action

This is the report showing use case workshop results and pathway to 30M + minutes of demand

Amazingly Simple Graphic Design Software – Canva

It appears that none of our work or reports were shared with the larger community.

We have been here not just to deliver workshops and onboard projects and companies but to build alongside the network.

Most of our proposal was to work immediately with broadcast production, conversational software for an AI reader, and to help pioneer a lip sync dubbing pipeline directly with a feature film to provide real use cases for the network. All of these projects have fixed industry deadlines and present unique opportunities to the network. They can’t be simply rescheduled. And in our opinion it doesn’t make sense to bring pipelines to the network if there is no strategy for demand which is what we worked to identify. The partners listed are top industry veterans and together bring access to a wide network that has the potential to bring long-term demand. Many are still in the figuring out their AI strategy phase which we were aiming to be early too.

We have been here hoping to build alongside even talking with orchestrators about BYOC and with GWID to help bring people into the network.

The power in Livepeer is the decentralised compute but the models that are available currently are not attractive to professionals in entertainment industry which is an established $3 trillion industry.

There has been many conversations about a focus on additional pipelines other than daydream and this is exactly what the research and the partnerships that we were working to bring in entail.

We are a start-up which has been mostly boot strapped and we spent most of our time trying to work with Livepeer because we believe in what’s possible. But in its current state other transcoding there is no clear value prop or road to demand. Even the Fashion hack has less than 10 participants

The reason the ask was more is because the first stage was R&D rather than building just to build or generate short-term demand we were looking at ways to drive long term usage, which we’ve found.

However, because no one is building the things that are needed, we don’t see a way to bring people in unless we actively do it with our team and partners.

I’m sorry that our progress and work was not shared with you earlier it’s a bit disappointing as we worked really hard on everything and it feels like no one saw it

The videos combined have this amount between long form and shorts.

Here we cover the problem and opportunity around both the dubbing and the reader solution. In the first part we did all of the conceptual
work and also started building a prototype here is the Figma walk through

Figma

How to Actually use AI in the Entertainment Industry

Figma

And here are YT stats from last 28 days.