According to Woofun AI, Dataline officially established its strategic position in the field of AI proxy data infrastructure. The core driving force stems from Robinhood (HOOD.US) co-founder and CEO Vlad Tenev's participation endorsement. The project aims to resolve data trustworthiness bottlenecks faced by AI agents before performing on-chain operations, and reshape the interactive paradigm of the cryptographic market data layer by constructing deterministic data pipelines and confidence scoring mechanisms.
This funding event not only marks capital's recognition of the transfer of the value capture layer between AI and Web3, but also indicates that data quality will become a key variable in determining the actual application of AI agents. Dataline plans to launch a native token by the end of 2026 and publicly sell it through Uniswap Continuous Clearing Auction (CCA) on the Base Chain to establish a full-stack ecosystem closed loop from data source to execution layer.
This vertical integration strategy, from low-level data verification to top-level application orchestration, attempts to establish a unified standard interface in the fragmented encrypted data market to provide reliable infrastructure support for high-frequency transactions and autonomous execution.
On October 22, Dataline announced the completion of a $10.6 million seed round, co-led by Lattice and Samsung-backed Dispersion Capital. A rich lineup of participants includes Symbolic Capital, Coinbase Ventures (COIN.US), Anagram, and celebrities such as Robinhood co-founder and CEO Vlad Tenev and Venmo (PYPL.US) co-founder.
In addition, former CFTC leaders and founders or partners from Anchorage Digital, NEA, General Catalyst, Jump Capital, FalconX, etc. also participated in this round of investment.
It is worth noting that despite the sizeable financing scale and deep funding background, the project has maintained a high degree of independence in terms of control structure. The financing was settled in April of this year, using a SAFE structure with token warrants, and the valuation was not disclosed to the outside world. More importantly, this round of financing did not transfer board seats, observer seats, or advisory titles to any investors, and the founders team fully retained control over the company's decision-making and governance.
This practice of refusing to cede governance in early financing is rare in the crypto primary market where top venture capitalists gather, showing the founding team's firm control over the long-term vision of the project. Woofun AI collates data and shows that this strong control structure helps the project maintain strategic consistency in the rapidly changing AI data circuit and avoid deviating from the core technology route due to short-term capital pressure.
Nuo Shi, co-founder of the agreement, clearly stated that the funds raised in this round will be mainly used in three major directions: expanding the coverage of data sources, strengthening the robustness of data verification mechanisms, and expanding into emerging fields such as robotics and AI training data. The project plans to launch its native token by the end of 2026 and has chosen to publicly sell tokens through the Uniswap Continuous Clearing Auction (CCA) mechanism on the Base Chain. The details of the offering are expected to be announced next week.
Dataline's predecessor, Tearline, started in 2025, when it officially entered the public eye after winning the BNB Chain AI Hackathon, and the name change was completed in May of this year. The evolution of the brand from Tearline to Dataline reflects the project's strategic intention to transform from consumer-grade applications to the infrastructure layer. Tearline initially focused on consumer-grade AI chatbots (ChatPilot) and automated scripts, but in the current primary crypto market, simple TG bots and automated scripts are facing the double pressure of aesthetic fatigue and peaking valuations.
However, the underlying data assets it has accumulated during operation provided a unique competitive advantage for the subsequent transformation into an AI data layer infrastructure. This strategic adjustment, which extends from the application layer down to the data layer, enables Dataline to use existing user behavior data and interaction records to optimize its data verification model and confidence scoring algorithm, thereby establishing differentiation barriers in the competitive data protocol circuit.
According to official information, the project team believes that the core bottleneck in whether AI agents can actually act autonomously is not the big model's reasoning ability, but the quality and trustworthiness of the data. Existing crypto market data is highly fragmented: exchanges, perpetual contracts, predictive markets, on-chain status, social signals, etc. each have different format specifications and update frequencies, making it difficult for AI agents to determine the freshness, multi-source consistency, and anomalous value identification when calling data. Dataline aims to provide a single integration point, return on-chain and off-chain data sources after normalizing them, and attach a confidence score to each response, so agents can decide whether to perform an operation based on this rather than blindly relying on a single price source. Nuo Shi pointed out that although there are data-related agreements such as Chainlink, Pyth, The Graph, and Grass on the market, Dataline is designed natively for AI agents and focuses on quality judgment and agent-friendly output formats before execution.
Currently, about 50 data sources have been launched on the platform, and the first AI Data Launch partners include Hyperliquid, Polymarket, Kalshi, Base, Robinhood Chain, BNB Chain, Sui, and TON. Developers building on Base and Robinhood Chain can access the service for free. According to the white paper snapshot data, the project has supported more than 3.4 million on-chain transactions, the success rate of proxy tasks has reached 96.4%, the median query delay is as low as 87 ms, and the cost of a single query is only 0.0001 USDC.
Although these figures are officially disclosed by the project party, and the actual operating indicators still need to be further verified in conjunction with public dashboards, their performance indicators have shown potential competitiveness in high-frequency trading scenarios.
It is worth noting that this round of luxury investment lineup betting reflects a shift in the 'value capture layer' on the crypto and AI cross-circuit. In the past two years, the AI Tier 1 market has concentrated heavily on the underlying computing power network and large-scale model training, but as AI Agents enter the actual stage of high-frequency trading, cross-chain arbitrage, and independent execution on the chain, the real bottleneck limiting the actual application of the agent economy, it has shifted from 'model reasoning capability' to 'data trustworthiness before execution'. Traditional oracles, such as Chainlink and Pyth, solve the core problem of a single price feed for smart contracts at the moment of on-chain settlement (on-chain settlement); in high-frequency multi-market games dominated by AI agents, agents need to simultaneously cross perpetual contracts (such as Hyperliquid) and prediction markets (such as Polymarket and Kalshi) to perform millisecond price comparison and arbitrage.
Once there is no uniform standardized data cleaning and anomaly filtering, a single dirty data or delayed signal will trigger serial liquidation of proxy policies. Dataline's differentiating advantage is its deterministic data pipeline design: all requests, whether from natural language or structured APIs, go through the same path — the intent analysis phase breaks down the query into markets, time ranges, and caller constraints; the routing phase selects appropriate data sources and supports multi-source fallback; the normalization phase unifies identifiers, units, decimal places and timestamps; the aggregation stage matches, filters exceptions, and preserves traceability information; the final output stage returns a structured response with quality metadata, which can be delivered via REST, MCP, etc.
This end-to-end deterministic processing process ensures that AI agents can obtain consistent and reliable data input in complex market environments.
The confidence score is Dataline's main differentiation design. Each response comes with a versioned score of 0 to 100, which takes into account multi-source consistency, data freshness, coverage, and source reliability. This score is the product of a fixed formula. It does not guarantee absolute accuracy, but rather gives agents a basis for evaluating risk before execution. For example, when stablecoin quotes deviate, the system will simultaneously return the anchored stability context and cross-market abnormal signals to help differentiate between potential arbitrage opportunities and de-anchoring risks. Quality assurance is achieved through an economic mechanism: data providers are required to pledge native tokens as a guarantee when submitting data. In the open data task network, the requester (requester) issues a reward task, and the fulfillment node must pledge DLINE as a quality guarantee. If poor quality data is submitted, the penalty mechanism will be triggered.
Forfeited tokens will be added to the original reward pool on the demand side as an additional bonus. Providers of regular data sources are required to continuously lock in tokens as a quality commitment. It is worth noting that Dataline is not a single data API protocol, but rather a four-layer full-stack matrix of 'data, orchestration, execution, and interaction' extended upstream by the application layer products of the predecessor Tearline. On top of the underlying Dataline Data, the protocol acts as a control plane through FlowAgent based on the LangGraph architecture, responsible for multi-step planning and confidence risk control and interception across complex DeFi strategies such as DEX arbitrage and revolving loans.
In the face of fragmented web terminals that have not yet opened an API, GhostDriver, the browser's native execution engine, can directly recognize the DOM structure and perform on-chain interactions on its behalf through the Chrome plug-in (official disclosure has accumulated over 10,000 users and 400,000 requests); the top-level ChatPilot acts as a traffic gripper on the Telegram side to provide natural language transactions and signage services for end users. The platform supports MCP native integration, making it easy to connect to existing AI development tools. Payment methods include subscription-based, x402 USDC pay-per-use payments on Base, and streaming micropayments on Tempo. Free credits are available for initial testing.
According to the white paper, the total supply of native tokens is 1 billion. The distribution structure is: 33% community rewards (330 million units), 30% ecology and foundation (300 million cards), Launchpad and liquidity 13% (130 million units), team 12% (120 million cards), and investors 12% (120 million cards). Approximately 21.5% of the supply is unlocked at TGE, and all are expected to be unlocked within 48 months. The main uses of tokens include online payments and data access. Users can lock tokens to continuously access data sources instead of traditional subscriptions. The data provider's pledge and forfeiture mechanism also relies on tokens to implement.
Details of the public sale have not been announced, including fund-raising targets and sales volume. The project party stated that it hopes to complete it on Base through Uniswap CCA and automatically inject initial liquidity into Uniswap v4 after the auction ends. The official white paper revealed that after the TGE token is completed, the data mission network and DLINE staking function will be opened. This is followed by the ZK certification layer and machine payment function. According to public information, the project was driven by core members such as co-founder Nuo Shi. The team is mainly distributed in Europe and North America. The exact number and full background were not further disclosed. The project was operated under the name Tearline before the change. Tearline initially focused on consumer-grade AI Chatbot (ChatPilot) and automated scripts.
However, in the 2026 crypto primary market, simple TG Bots and automated scripts are already aesthetically fatigued and have peaked in valuation. However, the underlying data it retains may show some value in the subsequent wave of AI.