Transcoder Campaign: SpeedyBird

Hi there!

We’re a new Orchestrator landing in the past two weeks. Our high-speed systems are geographically distributed and deployed in a highly automated fashion. This ensures we can move rapidly as new capabilities evolve and other opportunities emerge.

We did participate in the last few weekly water cooler calls as well as other community events. Our Orchestrator is backed by decades of experience building and running highly secure, scalable, and on-demand services in traditional data centers (on-prem) as well is in the cloud (AWS, Azure, GCP). All of these services were delivered under stringent regulatory demands. This means if you decide to stake with us, you are guaranteed a reliable partner that delivers under even the most demanding environment. You can find us right now at position 14 in the overall leader board.

Feel free to contact us via Discord! Be well my friends.

SpeedyBird (0xDef1C70578B2B5e8589a42e26980687fc5153079)

3 Likes

Two Years of Innovation: Update on SpeedyBird’s Livepeer Journey

Greetings, Livepeer community!

It’s been almost two years since my very first announcement, when SpeedyBird landed on the scene with a globally distributed, high-speed Orchestrator. Back then, we emphasized our commitment to enterprise-grade security, scalability, and reliability—values that continue to guide the work today. A lot has happened since those early days, so I’d like to share some of the most impactful projects I’ve been privileged to lead or contribute to within the Livepeer ecosystem. As always, you can reach me directly via Discord or replying here. Happy to discuss opportunities, node operations, and delegation structures that support our continued work.

1. Real-Time AI Video: Cascade

In partnership with Livepeer Inc., SpeedyBird developed and tested key components for the launch of the real-time video AI platform under very tight timelines with rapidly shifting requirements. By contributing to the Stream Manager aspects of this platform, users can create workflows (with ComfyUI or other inference models like StreamDiffusion) and gain access to:

  • Live Streams with AI: Processing real-time inference jobs to augment video content on-the-fly.
  • Simple Point and Click Pipelines: Offering a graphical environment for creating and managing AI-powered media services without the need for deep technical expertise. Essentially making this highly accessible to streamers worldwide.

2. Livepeer.Cloud SPE Initiatives

Over the past year, I served as an active member of the Livepeer.Cloud SPE, collaborating on major projects that have helped extend and simplify the network’s capabilities. We focused on:

  • Free-to-Use Gateway: Facilitated open access to the network by securing funding for jobs and deploying the public Livepeer.Cloud Gateway, enabling free usage for all. View it’s stats here → Grafana
  • Integration with Owncast: Contributing to the successful integration of Livepeer into Owncast, bringing decentralized live streaming to a wider audience.
  • Gateway Node Documentation: Contributed to clear, comprehensive documentation, which reduced onboarding friction and accelerated network growth from 8 Gateway Nodes to over 35.
  • Advanced Metrics & AI Leaderboard: Enhanced the Livpeer Explorer Leaderboard and Data API to track transcoding and AI job performance, resulting in real-time AI performance metrics and expanded transparency for Orchestrators.
  • Decentralized Metrics Architecture: Provided an architecture for decentralized metrics, which Livepeer Inc. leveraged in its Kafka-based solution. This powers insights into AI inference work beyond the leaderboard for deep insights. A potential next step for the SPE is to push further toward a fully decentralized approach, ensuring robust data flows without single points of control.

Thanks to these contributions, we continue to see Orchestrators benefit directly and have enabled the growth in Gateways across the network, illustrating how collaborative innovation enriches the entire ecosystem.


3. Video Miner Pool Development

As part of the Video Miner project, my collaborators and I built a transcoding (and soon AI-enabled) pool to offer an alternative, stable resource for operators. Key highlights include:

  • User-Friendly Web Interface: Allowing pool participants to monitor metrics and performance with ease.
  • Simplified Onboarding: Providing Windows and Linux installers that eliminated much of the complexity for new Transcoders.
  • Infrastructure Diversification: Expanding the ecosystem beyond a single, well-managed pool, ensuring more robust options for those looking to participate in transcoding.

This pool helped many new and experienced operators find a reliable, easy-to-use path into the Livepeer ecosystem.


4. Advanced Monitoring & Alerting Stack

To streamline operational insights, I developed a fully containerized Docker Compose project that bundles Grafana, Prometheus, and Loki:

  • Prebuilt Dashboards & Alerts: Allowing operators to quickly visualize performance and receive real-time notifications.
  • Extensive Documentation: Helping both new and seasoned operators tailor configurations and easily add custom dashboards.
  • Community-Centric Tools: Ensuring that the orchestration layer remains transparent, trustworthy, and easier to maintain.

This stack (available here) saves time for node operators and fosters best practices across the network. More on this is detailed in the blog post here.


Additional Ongoing Contributions

Beyond the major initiatives outlined, I regularly contribute through testing, documentation reviews, and hands-on engagement in the Livepeer Discord and direct engagement with community members. This day-to-day collaboration helps ensure the advancement of the community and network at large. I am honored to work with this community!


Looking Ahead

From free gateways to AI-enhanced streaming, from pool management to advanced monitoring, my overarching mission remains the same: to build, collaborate, and push the boundaries of what Livepeer can do. Throughout these initiatives, I’ve witnessed the network’s transformation into a thriving ecosystem filled with ambitious projects and passionate contributors.

If you’ve been following my journey since that first announcement, you’ve likely noticed a steadfast commitment to reliability, innovation, and community-driven growth. As we continue down this path—embracing new technologies, onboarding more operators, and driving network improvements—I invite everyone to keep an eye on the road ahead. There’s still much to build, and I’m here for the long haul.

Thank you for your continued support and collaboration. I look forward to driving even more value in the months and years to come!


— SpeedyBird

Dedicated to delivering reliable, innovative services in Livepeer—and always open to new ideas and collaborations.

10 Likes

SpeedyBird Orchestrator Update: Scaling GPU Power & Network Contributions

Greetings Livepeer community,

SpeedyBird has been building in Livepeer for three years now, and I’m excited to share the next stage of growth. Delegators often ask how additional stake supports the network. The answer is simple:

  1. More GPUs online to increase capacity for transcoding and AI jobs
  2. More time invested into development, research, and ecosystem contributions

This dual focus ensures that delegating to SpeedyBird strengthens both the infrastructure of the network today and the innovation needed for tomorrow. Additionally, this powers contributions to the Cloud SPE (see below for more details).

As stake grows, SpeedyBird continues to expand GPU capacity while dedicating more time to Livepeer software development, research, and Cloud SPE initiatives that strengthen the broader ecosystem.


Current Operations

  • 6 GPUs Globally Distributed:
    • 1× RTX 5090
    • 1× RTX 4090
    • 1× RTX 3090
    • 3× GTX 1080s
  • Enterprise-grade East Coast server (owned): Datacenter-level hardware → Supermicro motherboard, 32-core AMD EPYC 7302, 128GB ECC RAM, redundant Intel P5500 NVMe RAID, solar-backed + 1Gbps fiber.
  • Additional datacenter nodes provide uptime, redundancy, and global coverage.

Why Delegators Should Choose SpeedyBird

  • Active & Committed: Currently ranked #44 by stake, yet consistently one of the most engaged and contributing operators — running the latest updates, deploying new GPUs, and supporting Cloud SPE initiatives month after month.
  • Proven Contributions: From building the free Cloud Gateway to contributing in the early stages of Live AI Video (Showcase/Daydream), the AI Leaderboard, advanced Metrics stack, and Cloud SPE tooling, SpeedyBird has consistently delivered value beyond node operations.
  • Enterprise Reliability: Redundant design, enterprise hardware, and sustainable energy.
  • Global Datacenter Presence: Capacity across multiple regions.
  • Long-Term Commitment: 3 years operating with a strong focus on ecosystem growth. Consistently active in community calls and Discord with impeccable uptime and reward calls.

Cloud SPE Focus

Beyond individual operations, I’m also part of the Cloud SPE, where we’re pushing forward several ecosystem-wide initiatives:

  • Public gateway with free access
  • GPU pool development & fair revenue sharing
  • Job tester with AI Support (including live video)
  • Robust demos, docs, and public APIs
  • Open metrics and testing capabilities

Recent Cloud SPE Technical Updates

In recent weeks, we’ve focused on expanding the Cloud SPE’s ability to handle Live Video AI jobs, including:

  • Enhancing the AI Job Tester to support live video stream inputs and playback quality detection.
  • Adding scoring logic for new job types that were previously unsupported in the testing framework.
  • Integrating MediaMTX and managing multiple service URIs not yet supported in the onchain registry.
  • This work extends prior completed (and funded) proposals and required significant development and testing beyond simple configuration updates — all completed without additional funding. In fact the funding for these services ran out months ago.

:book: Read the full Cloud SPE update here


:rocket: Looking Ahead

Delegating to SpeedyBird doesn’t just put GPUs online — it ensures continued innovation, transparency, and ecosystem growth. With your support, I’ll continue scaling both infrastructure and development to keep Livepeer at the forefront of AI-powered video.

Thank you for your trust and continued delegation.

SpeedyBird

1 Like