James / 0Impact EcoCredit Agent
James shared that the EcoCredit agent platform is live. The system runs on 0G decentralized compute, which James described as 35% cheaper than traditional cloud compute and designed to support carbon-offset-aligned AI infrastructure. Agents are fully connected to EcoBridge and can be configured to actively drive eco-credit sales: each agent generates a personal referral URL, and any offsets or donations routed through that link earn the agent’s owner RWI tokens, which are auto-staked to the account. Accumulated RWI unlocks the ability to mint additional agents and activate new skills — creating a flywheel where agents that demonstrably drive impact earn the capacity to do more.
Agents are programmable via a toggleable skills interface. Current integrations include X, Discord, and Telegram; email is coming once a Cloudflare infrastructure update lands. Agents can also accept MCP hookups and custom-trained skills. Each agent can be minted as an INFT, meaning an agent with a proven track record of revenue generation can be sold as a small business. A skills marketplace is in development that would allow agents to sell their skills to other agents. James offered compute credits to anyone on the call willing to help stress-test the skill set before the platform’s next finalization pass. Gregory noted the platform could be useful for the Regen Foundation as a grants agent, and James confirmed 0Impact has a fairly robust grants agent available.
Brandon asked about customer journey mapping and whether the agent stack could be applied to construction site ecological tracking; James confirmed it can, and Gregory noted the sensor intake work being built for Regen Registry would generalize to any sensor context including construction materials and site data.
Christian / Keeling Curve Internship & Regen Compute
Christian announced the formal start of a 14-week internship program with two developers sourced through the Keeling Curve Prize program. The agreement was signed the previous day. The original plan had been a free-form build-on-Regen-Compute format, but after a session with Gregory and the Gaia AI team, the interns were redirected toward building sensor intake SDKs and APIs for Regen Registry. Christian used Claude to cross-reference the interns’ CVs with the technical requirements and found the match strong: one intern had previously led SDK development for a race car team’s multi-sensor telemetry system, translating directly to the data ingestion pipeline work. Mentorship will come from Darren and Samu on the Gaia AI side, with Gregory providing technical direction on the registry architecture. The interns are expected to present progress to this working group over the course of the program.
On Regen Compute, Christian noted there has been no active marketing push and the platform sits at 35 subscriptions. He had calls this past week with both Klima and Guardians of Earth. Guardians of Earth (represented by Andrew) is interested in creating a “bio unit” credit class on Regen Ledger and listing it as a Regen Compute offering, with a goal of reaching 500–1,000 users as a threshold to approach larger AI companies about integrating ecological offsetting as a core feature. Klima is focused on the corporate sales side while Regen handles procurement and project developer relationships; Christian sees these as complementary and is exploring a compute collaboration using the existing CFC bridge infrastructure. He also updated James that the CFC reconciliation — previously blocked on finding the right process documentation — is now unblocked: a video call is scheduled this week to complete the accounting, ahead of James’s upcoming Solana Policy Institute call.
Gregory Landua / Claims Engine Architecture & Registry Update
Gregory walked through the architecture that the interns will be building against, framing it as a standardized system for data ingestion, data processing, and claims — designed to work at a level of abstraction above eco-crediting specifically. The core approach uses a semantic graph schema (aligned with Hyperserts and JSON-LD/COI federation patterns) to guarantee determinism in agentic outputs and prevent hallucination when models navigate ecological data. The sensor intake layer is designed to handle heterogeneous inputs: iNaturalist API, Otter transcripts, Twitter, Planet Labs satellite data, ESA Sentinel, and custom feeds. By solving the hardest evidence-verification problem (eco-crediting), the stack can be ratcheted back to business operations dashboards, construction site tracking, or any other multi-sensor intelligence application.
On the registry: Gregory confirmed that Regen Registry is formally beginning the transition to operate under Regen Foundation, the 501c3. R&D Inc. will continue as the registry operator of record and contractor, but the foundation will be the owner and governing body through its independent board process. Alongside the registry, Regen Commons and Austin’s River Computer project will each become independent programs under the foundation, each with its own budget but deeply complementary missions. Gregory described the registry as a strong philanthropic fundraising narrative — a concrete, tangible digital public good that land stewards and organizations rely on — and plans to share a foundation pitch with this working group within the next month to gather feedback and begin mobilizing the network toward a capital campaign. He noted that anyone who encounters potential philanthropic donors in their networks should reach out once the pitch is circulated, as contributions would be tax-deductible.
Gregory also demoed Planet AI’s beta satellite interface, which provides daily full-earth imagery and a conversational analysis layer. He ran a test on Wilbur Hot Springs — a site where Regen had previously assessed beaver dam analog projects — querying whether the landscape was rehydrating. He also pulled an urban sprawl analysis on Almaty live on the call, showing 42% population growth over the last decade.
Max Semenchuk / Regen Bank, Alpha Earth & Base Migration
Max reported that Regen Bank is preparing to apply to upcoming Celo funding rounds via Presentee. A positive signal came from Max Zalevsky, a land steward from Ukraine whose iNaturalist credits are listed on the bank: after a Crypto Altruist article about the platform, Zalevsky promoted it independently and is now co-applying for the Celo grant. Max framed this as validation of the cooperative model — project owners who benefit from credit sales become active promoters of the platform. He identified local currency acceptance as a near-term requirement for deeper Kazakhstan engagement, noting that an acquiring agreement would be needed to accept local currency and convert it for on-chain use.
Max demonstrated a visualization he built using Google Earth AI (via the Earth Engine SDK), showing air pollution heat maps, greenness index, and temperature change data for Almaty across multiple years. He framed this as a potential MRV data layer — grid-based spatial data where reforestation or other ecological projects could be tracked at the neighborhood level using publicly available satellite datasets. Gregory noted that Planet AI’s beta, which he was simultaneously testing, takes a similar approach but with a conversational interface and daily imagery updates, and pointed to the Regen eco-weaving channel as a place where access had been shared.
On the Base migration question — raised by Klima’s Alex writing a forthcoming analysis — Max noted that the Celo-based Regen Bank contracts are already EVM-compatible, making a technical port to Base straightforward. His hesitation is the added user complexity of managing credits across two chains. His threshold for doing it: a committed demand signal rather than just infrastructure funding, on the order of a $300K credit purchase commitment from Base ecosystem parties. Gregory echoed this, adding that ideally any migration would come with both budget to cover development and guaranteed demand.