Welcome to the next crypto level
← Nexorena

Blockchain

Lightchain AI announces a test of agents billed per request

The roadmap released on November 19 calls for named agents on the test network, payment per call in LCAI, and a demonstration on December 12. An external review is to precede the mainnet deployment decision.

Nexorena Editorial Team

Lightchain AI announces a test of agents billed per request

Named agents on the test network

Lightchain AI presented a roadmap on November 19, 2026, that runs through January 2027. The project wants applications to be able to find an artificial intelligence service by name, send it a request, and pay per call. The published schedule describes testing stages. It provides for a decision on whether to launch the agent contracts after testing, without setting a mainnet launch date.

The first stage is planned for November. Lightchain intends to create an agent called atlas.lcai on the test network and open a gateway that applications can use to find agents by name. At least two operators are to run the Llama 3 8B model there. The verification service is to rerun every test job. On mainnet, the project says it currently reruns only a sample of jobs. The test network results would therefore provide only part of the evidence about how the service would work in production.

A payment linked to each call

In December, Lightchain wants to test an LCAI payment for each call to an agent. The proposed system links the payment to a specific request. For an application that uses several services, this would make it possible to match each response to the call that triggered it and the corresponding payment. The roadmap also provides for one agent to pass a task to another. The team announced a public demonstration for December 12, with a call from Claude Code and a payment initiated by a script.

Two different checks are involved. Matching a response to a job verifies its origin within the system. Rerunning a job checks the work produced. A receipt tied to a request is therefore insufficient to establish the quality of the response. To assess the system, observers will need to see which jobs were checked, how discrepancies are handled, and which results can be reproduced from the records the project plans to publish.

A launch decision after external review

Lightchain plans to send the agent contracts for external review in December. In January, the project aims to publish the report, extend the verification service to every model it accepts, and rehearse deployment on the test network. The release does not say the report is already available. The decision on whether to launch the contracts on mainnet is to be recorded after the rehearsal.

Each stage is to leave references that can be viewed in the test explorer and be covered in a monthly report comparing deliveries with targets. These records should make it possible to track the first agent, per-call payment, and deployment preparations separately. The published dates, including the demonstration date, remain targets set by Lightchain.

Article sources

Written with love byEthan BrooksMeet this editorial pen name ↗

Explorer Nexorena