Perle

Perle is a sovereign AI data layer that uses Solana's blockchain infrastructure to make the provenance of AI training data transparent and auditable. Rather than sourcing annotations from anonymous crowdsourced pools, Perle routes tasks to vetted domain specialists—physicians, lawyers, robotics engineers, security professionals—and records every contribution, verification, and quality signal on-chain. The result is a tamper-resistant history of who labeled what, enabling enterprises and AI labs to satisfy increasingly stringent regulatory and governance requirements around data supply chains.

The project operates under Perle Labs and was founded in early 2024, previously operating under the name Kiva AI before rebranding. CEO Ahmed Rashad leads a team that includes veterans from Scale AI, Meta, MIT, and Berkeley.

The Data Quality Problem

AI labs training large language models and multimodal systems face a deepening quality crisis: most annotation pipelines rely on anonymous gig workers whose qualifications cannot be verified and whose work histories cannot be inspected. Perle's premise is that specialized tasks—clinical reasoning, legal text review, robotic manipulation labeling—require domain expertise that generalist crowd platforms cannot supply reliably.

Perle describes its service offering as data acquisition, annotation, evaluation, and reinforcement learning from human feedback (RLHF) across multimodal formats. Rashad has cited internal benchmarking in which Perle's human-in-the-loop annotation outperformed Amazon Rekognition by over 70%, though this figure has not been independently verified.

How the Platform Works

Perle operates as a hybrid on-chain and off-chain system. The economically substantive work—task routing, annotation, review—happens off-chain through Perle Labs Creator, a two-sided marketplace that connects organizations requiring training data with qualified contributors. Creator handles the full workflow cycle: task creation, expert assignment, multi-stage review, and final delivery.

The Solana blockchain handles the attribution, settlement, and governance layer. Every contribution and verification result is written to the chain, creating an immutable audit trail that enterprises and governments can inspect for regulatory compliance. Smart contracts escrow contributor rewards, releasing payment upon verified task completion rather than leaving compensation subject to platform discretion.

Identity verification mechanisms, including Proof-of-Personhood, are used to filter out bots and ensure that task attribution maps to real human contributors with trackable quality histories. Contributors build verifiable on-chain reputations over time, which influences task allocation priority within the system.

Expert Network

Perle has assembled what it describes as a network of over one million annotators, including more than 2,500 physicians and 530 domain specialists. The emphasis on credentialed professionals distinguishes the platform from general-purpose annotation marketplaces and targets the segment of AI development where low-quality labeling poses the highest downstream risk—clinical AI, legal document processing, defense-adjacent applications, and complex robotics scenarios.

Contributors earn PRL tokens for completing high-accuracy tasks. The reward structure is governed by smart contracts, meaning payment terms are set programmatically rather than unilaterally by the platform.

Solana Integration

Perle formally announced its expansion into the Solana ecosystem on September 23, 2025. The choice of Solana was stated as practical: the network's high throughput and near-instant finality make it viable to record task-level attribution events at the scale Perle's workflow generates, without the settlement delays or gas cost unpredictability that would make Ethereum-based record-keeping operationally impractical for a data coordination platform.

Rishin Sharma, Solana Foundation Head of AI, commented at the time that Perle's permissionless platform approach aligned with the Foundation's vision for blockchain advancing AI development.

PRL is deployed as an SPL token on Solana with the contract address PERLEQKUNUp1dgFZ8EvyXHdN9d6ZQqfGxALDvfs6pDs.

Funding

Perle has raised approximately $17.5 million across two rounds. A pre-seed round of $8.5 million closed in October 2024. A $9 million seed round, led by Framework Ventures, was announced on August 7, 2025, with CoinFund, Protagonist, HashKey Capital, and Peer VC participating. The seed funding was positioned as the launch capital for Perle Labs, the crypto-native layer of the business.

PRL Token

The PRL token has a fixed supply of one billion tokens, with no provision for future minting. Token allocation breaks down as follows: community at 37.5%, investors at 27.66%, ecosystem at 17.84%, and team at 17%.

Vesting is structured to back-load liquidity. The team and investor allocations receive 0% at the Token Generation Event, with 25% unlocking at month twelve and the remaining 75% distributed linearly over the following 36 months, reaching full unlock at month 48. The ecosystem allocation releases 10% at TGE with the remainder distributed linearly over 48 months. Community allocation releases 7.5% at TGE with the balance distributed linearly over 36 months. Day-one circulating supply is 17.5% of total.

The TGE occurred on March 25, 2026. A Season 1 airdrop campaign preceded it, during which the Perle network completed over 1.7 million tasks and distributed 330 million points to participants, who then converted points into PRL allocations at launch.

Within the platform, PRL serves as the coordination token: contributors earn it for completing tasks above accuracy thresholds, enterprises use it to access data pipelines and prioritize task queues, and governance participants use it to influence protocol parameters. The token is not required for simple consumption of the platform's services, but it aligns the economic interests of contributors and enterprise customers within a shared incentive framework.

Positioning in the AI Data Market

Perle is entering a market that includes both Web2 incumbents—Scale AI, Appen, Sama—and a growing set of Web3-native competitors. Its differentiation rests on three claims: that its expert network is more credentialed than general-purpose annotation platforms, that on-chain attribution provides auditability those platforms cannot match, and that blockchain-based escrow makes contributor compensation more reliable. The project targets enterprise customers and frontier AI labs that face increasing scrutiny over data governance, particularly in regulated sectors where the traceability of training inputs has compliance implications.

The platform's commercial traction, task volume at scale, and enterprise customer adoption remain to be established publicly as the project moves past its TGE into live operations.