What is Livepeer?
👋 New to this? Just start reading at the top — it begins in plain English and gets more detailed as you scroll. Jump to any level:
🟢 The simple version
Plain English — no jargon. Start here.
Livepeer is a decentralised video transcoding network — when a video is uploaded or streamed, Livepeer distributes the computational work of converting it into different formats and resolutions across a network of GPU-equipped computers, paying operators in LPT rather than routing everything through expensive centralised cloud services.
The video transcoding problem
When you upload a video to YouTube or stream live on Twitch, the video must be transcoded — converted from its original format into dozens of different versions: 1080p, 720p, 480p, 360p, different codecs, different bitrates. This ensures every viewer gets a version appropriate for their device and connection speed. Transcoding is enormously compute-intensive — YouTube alone processes hundreds of hours of video per minute.
This transcoding work is done on centralised cloud infrastructure (Amazon, Google, Microsoft). It is expensive: video platforms spend significant portions of their operating budgets on transcoding infrastructure. Livepeer proposes replacing this centralised infrastructure with a network of GPU-equipped node operators who compete to perform transcoding jobs, paid in ETH by the requesting platform and earning LPT as additional rewards.
How Livepeer works — orchestrators and delegators
Livepeer has two participant types. Orchestrators are node operators — they run Livepeer software on GPU-equipped machines, accept transcoding jobs, complete them, and receive payment. Delegators are LPT holders who do not run nodes but stake their LPT with orchestrators they trust, sharing in the orchestrator's rewards proportionally. This delegation model is similar to DPoS — LPT holders collectively select and economically support the most reliable orchestrators.
What LPT is used for
LPT is the staking token for the Livepeer network. Orchestrators must stake LPT to participate — more stake = more work allocation. Delegators stake LPT with orchestrators to earn a share of rewards. LPT is inflationary (new tokens distributed to active participants) — the inflation rate decreases as more LPT is staked, incentivising participation.
Is it legal in India?
Yes. LPT is a VDA under Indian law. 30% tax and 1% TDS apply. See India regulation.
🟡 A bit more detail
For when you want to go a little deeper.
The Livepeer network — real video processing
Livepeer is not theoretical — it processes real video for real customers. Video platforms including Manifestoo, Bonfire, and dozens of others use Livepeer's open-source stack and decentralised network to reduce their transcoding costs. The Livepeer Studio (previously called Livepeer.com) provides an API that developers can call to submit video transcoding jobs to the decentralised network — abstracting the blockchain complexity entirely. A developer adds a single API endpoint and the network handles routing, orchestrator selection, quality verification, and payment.
Livepeer is used by video streaming startups, decentralised social platforms (Lens Protocol uses Livepeer for video), and Web3 gaming platforms requiring on-chain video. The network processes millions of minutes of video monthly. AI video generation platforms have begun using Livepeer for GPU compute beyond transcoding — generating video from prompts requires similar GPU resources to transcoding, and Livepeer's decentralised GPU network is being extended to these workloads. This AI compute expansion mirrors Theta Network's parallel evolution.
The probabilistic micropayment system
Video transcoding involves thousands of small jobs (individual video segments). Paying for each on Ethereum mainnet would cost more in gas than the job is worth. Livepeer uses a probabilistic micropayment system: instead of paying for every job, the client sends "winning tickets" — cryptographic lottery tickets where each ticket has a small probability of being worth a full payment. Statistically, over many tickets, the expected payout equals the correct fee. Orchestrators only redeem winning tickets on-chain (rare events), dramatically reducing transaction costs.
Key Livepeer metrics: minutes of video transcoded per day, number of active orchestrators, ETH fee revenue earned by the network, and LPT staking rate. The core metric is actual video processed — not token price or trading volume. An orchestrator's performance (uptime, transcoding accuracy, latency) determines their work allocation and earnings. Live data: CoinGecko · Livepeer Explorer.
🟣 The full technical picture
For the technically curious.
The Livepeer whitepaper: probabilistic micropayments
The Livepeer whitepaper (Caplan & Zheng, 2018) introduces the "Probabilistic Micropayment" scheme formally. The scheme works as follows: a client opens a deposit with a Ticket Broker smart contract. When requesting a transcoding job, the client sends a signed ticket specifying a face value V, a win probability P, and the recipient (orchestrator). The orchestrator verifies the ticket signature and the client's deposit covers the expected value (V × P). With probability P, the ticket is "winning" and the orchestrator can redeem it for V on-chain. With probability (1-P), the ticket has no value but costs the orchestrator nothing to verify. Over N jobs, the expected payout is N × V × P, matching the correct total payment for N jobs. The on-chain overhead is proportional to 1/P rather than N — if P = 0.01, only 1 in 100 tickets requires on-chain settlement.
Source: Caplan, D. & Zheng, E. (2018). Livepeer Whitepaper. github.com/livepeer/wiki · Livepeer docs: docs.livepeer.org
Verification and the truebit challenge
How does the network verify that an orchestrator actually performed the transcoding correctly without re-doing the work? Livepeer uses a challenge-response model. When a broadcaster submits a transcoding job, it keeps a sample of segments. After receiving transcoded output, the broadcaster can challenge a specific segment — submitting the original and asking for verification. Livepeer's verification mechanism uses a combination of deterministic transcoding (the same input should always produce the same output with given parameters) and spot-checking — randomly verifying a small subset of segments makes fraud statistically detectable and economically unprofitable.
LPT's inflation model: the Livepeer protocol targets 50% of all LPT being actively staked. When staking participation is below 50%, the inflation rate increases — incentivising more staking. When above 50%, the inflation rate decreases. This creates a self-balancing mechanism: participation is economically incentivised to converge toward the 50% target. The inflation rate has a maximum and minimum bound (both configurable by governance), preventing runaway inflation or deflation. LPT holders who do not stake are diluted by inflation — creating a "stake or lose value" dynamic that ensures network security is funded by active participation rather than passive holding.
Key facts
- Network function: Decentralised video transcoding
- Participants: Orchestrators (GPU operators) + Delegators (LPT stakers)
- Payment: ETH (per job) + LPT inflation (network rewards)
- Settlement: Probabilistic micropayments (Ticket Broker)
- Chain: Ethereum (Arbitrum for gas efficiency)
- LPT supply: ~30 million (inflationary, self-balancing)
- Founded: 2017 (Doug Petkanics, Eric Tang)