Arca Launches App for Payments Based on AI Interaction History

On November 11, 2026, the Brevis development team announced the public launch of Arca, an application for Windows and macOS that allows users to earn rewards for providing their history of working with artificial intelligence tools. The service is designed for users who utilize AI in programming, research, document preparation, and other professional tasks. After a closed testing period that began on October 13, registration is now available without invitations or prior approval.

Arca is based on the idea of using the history of AI interactions as a separate digital resource. While working with language models, users create sequences of queries and responses, correct errors, refine instructions, and verify results. These records may contain information about real work processes, which the service developers suggest submitting for subsequent evaluation and reward.

How Arca Works and What Payments Are Provided For

Arca is compatible with several popular artificial intelligence tools, including Claude Code, OpenAI Codex, Cursor, and Claude Desktop. After installation, the application detects session history already saved on the computer, including previously completed interactions with models. Users independently select the records they want to submit for evaluation, and the service offers an estimated range of possible rewards.

Before submitting data, Arca, according to the developers, automatically removes names, local file paths, account information, and access keys. After preliminary processing, the selected session is sent for evaluation, and users can confirm further consideration or decline the offer. In case of refusal, the uploaded copy is deleted, as claimed by the company, and additional anonymization is also provided on Arca's side.

Rewards are only credited for materials that pass verification. Developers clarify that very short AI interactions do not meet the service's requirements, and the payment amount depends on the content and quality of the submitted work. Approved sessions are paid weekly in USDC stablecoins, but fixed rates and guaranteed minimum reward amounts are not disclosed.

Additionally, users can connect a GitHub account to verify professional experience. According to Brevis, this may enhance the evaluation of submitted materials, although integration is not mandatory. Currently, the program supports certain Claude and GPT models, and developers plan to expand the list of available tools.

What Risks Are Associated with Transmitting AI Session History

The main concern when using such services is the confidentiality of the transmitted information. The history of interactions with artificial intelligence may include source code, work correspondence, internal company documents, and other information not intended for distribution. Even after automatic removal of personal data, materials may still contain fragments that allow determining their origin or reconstructing individual details.

Brevis states that Arca does not upload history in the background and allows users to independently select sessions to be transmitted. However, independent verification of anonymization mechanisms has not yet been presented, so the company's statements should not be interpreted as a guarantee of complete information protection. Before sending work materials, users also need to consider employer confidentiality conditions, contractual restrictions, and third-party rights.

Public Launch and Contest with a $10,000 Fund

Arca initially became available on October 13, 2026, as part of closed testing, where participants received invitations in small groups. As of November 11, the service is open for registration via email or Google account, and the application itself is distributed for free. Developers continue to enhance the session evaluation system and mechanisms for determining rewards for submitted materials.

Simultaneously, the team announced the Arca Launch Challenge promotional campaign with a total prize fund of $10,000, provided in addition to regular payments. According to the published rules, three participants with the highest results are to receive $1,000, $600, and $400, respectively, while the remaining $8,000 is distributed among other eligible participants proportionally to their results. However, the published company materials contain different end dates for the promotion, so participation deadlines require further clarification.

The public launch of Arca demonstrates the emergence of services offering users financial rewards for data generated during work with artificial intelligence. However, the practical value of this model will depend on the transparency of session evaluation, the stability of payments, and the developers' ability to ensure the protection of transmitted information. Independent data on average user earnings and the effectiveness of the provided confidentiality measures are still insufficient.