Mode Inc buys Trimbox and QR Code Reader, targeting everyday consented data for AI

Mode Inc has turned AI data collection into an acquisition funnel. The company just bought Trimbox, an inbox management app, and QR Code Reader, a code scanning app, expanding its portfolio to include more than 100 million monthly users. The point is not subtle: Mode wants everyday consumers to trade their data for cash or rewards, and it is scaling that model by buying the apps that already sit on people’s phones.

CEO Dan Novaes framed it clearly to Business Insider. He said Mode is “really about everyday consumers” rather than chasing pay-per-gig contractors from sites like Mechanical Turk and Outlier/Scale AI. In practice, that means Mode is built around data types people generate routinely, including uploading receipts from purchases at places like Amazon or Walmart, streaming data, and wearable device data. Those receipts, streams, and device signals are then used to help train AI systems, the same broad category of work that has made labeled and structured datasets a recurring bottleneck for developers.

This is also why the acquisitions matter beyond Mode’s internal strategy. Mode is not positioned as a direct competitor to the big dataset-and-training-services ecosystem, but it sits in the same ecosystem of startups that use paid labor to collect and label information. The source name-checks Scale AI, Mercor, and Handshake, which pay hundreds of thousands of contractors around the world to collect and label data used to improve products like self-driving cars and AI chat systems, including OpenAI and Meta’s chatbots.

Mode’s differentiator is incentive design and consent framing, not the existence of data work. The company pays consumers for their everyday data, which it characterizes as different from gig-focused AI training startups. Mode then adds more “niche apps” to its portfolio, each expected to pull in millions of monthly users. Novaes described the math in blunt terms: there are two paths to a billion monthly active users, either build the next Telegram or Twitter, or acquire 1,000 apps with a million monthly active users each. For boards and investors, that means Mode is treating distribution like the product, then using that reach to generate training signals.

The regulatory and legal backdrop is the other half of the thesis. Novaes said growing legal scrutiny around AI companies’ use of online content will spike demand for consent-based data collection. He cited “recent lawsuits against Anthropic and Perplexity” as examples of the pressure rising on how AI systems use content, even when the data is sourced through consumer activity rather than traditional licensing deals. That is the risk tradeoff that consent-driven models are trying to monetize: if AI labs expect more proof that data was collected with permission, companies that already operationalize user consent and payment can become default vendors.

Mode provided a concrete example of client requests to illustrate the shift toward consented, targeted collections. Novaes described an AI lab client that wanted people to submit forms that have handwriting on them, referencing examples like doctors’ notes or oil change receipts. The request, he said, was for millions of documents of that type. Mode’s response, according to the source, was to send users a single notice: “Get some samples.” This is how the business ties back to the acquisitions. Trimbox and QR Code Reader are not just consumer utilities; they are potential channels for harvesting structured, usable signals, from receipts to scanned codes, with a payment loop for users.

The capital story adds urgency too. Novaes said he founded Mode in 2019 so people could get paid for the time they spend on their devices and the data they generate. The company, according to the source, has handed out $1 billion in earnings, savings, and incentives. It has raised more than $80 million through crowdfunding, and Novaes said he intends to take the company public in the next two years. For executives, that timeline turns the acquisition spree into a governance question: how quickly Mode can turn app scale into high-quality, consented training datasets, while keeping compliance and partner demand aligned, before the market demands evidence.

Ultimately, this is a playbook other AI data builders may watch closely. If legal scrutiny continues to tighten and AI labs increasingly demand datasets that can be defended as consented, then distribution plus consent mechanics could become competitive advantage, not just ethical positioning. Mode’s current roadmap, as described in the source, is to build a portfolio of niche apps that collectively reach massive scale, and to keep acquiring when it can add both users and useful data capture. The second-order implication for investors and board members is straightforward: the winners may not only be the labs training models, but the companies that control the supply chain of permissioned data at scale.