Monthly Report - September 2025

General

News

TL;DR

September was all about moving Ratio1 from a fully‑functional platform to a production‑ready ecosystem that can power enterprise‑grade AI applications. Key highlights include:

Deeploy v2 released on mainnet and dAuth v2 in the works, bringing automated orchestration, job editing, invoice/burn‑report features and a completely automated escrow system. These upgrades build upon the simplified decentralized authentication flow introduced earlier, making node onboarding, SDK access and job lifecycle management frictionless and bringing web2 simplicity to web3.
The first full enterprise application built on Ratio1 went live. Partner Keysoft used J33VES to create AI assistants that transform plain‑language descriptions into database schemas and user interfaces, generate SQL‑driven reports and enforce policy‑grounded chatbots, all while keeping data encrypted in R1FS and using decentralized identities through dAuth.
RedMesh v1, a decentralized, distributed penetration‑testing framework, was released following last month's inception. It performs continuous security testing across edge nodes and is open‑sourced to encourage community participation. CerviSafe - a proof‑of‑concept for privacy‑preserving cervical‑cancer screening - demonstrates how Ratio1 can enable secure healthcare workflows using federated learning and homomorphic encryption. LabelCraft v0.5 public beta and Petware cross‑native/generic AI templates entered testing, expanding the AI‑for‑Everyone toolkit.
September was indeed the international expansion focus month - Ratio1 held extensive meetings with international consortiums (including Horizon Europe, Pathfinder and EIC) to embed decentralized AI in large‑scale healthcare and cybersecurity projects. Partnerships span more than 15 countries and involve tens of companies across academia, research institutes and industry.
Last but not least community growth & education - The AI‑for‑Everyone creation toolkit and LabelCraft saw major enhancements. Educational resources (e.g., TikTok tutorials and Ratio1 Courses) continue to be developed in order to draw new developers and non‑technical users into the ecosystem. NodeJS integration and template applications like Ratio1 Drive make it easier for developers in other ecosystems to build on Ratio1.

In depth

Enterprise application powered by J33VES (R1‑295)

First enterprise deployment: Keysoft’s production‑grade application built with the J33VES framework went live. The app includes three AI assistants: a schema and UI generator that translates plain‑English descriptions into relational database schemas and interface code, a report builder that produces SQL‑based dashboards from those schemas, and a policy‑grounded chatbot for compliant support.

Privacy by design: Each assistant runs on open‑weight models and neuro‑symbolic tools hosted on the decentralized network. User data and knowledge stay encrypted in R1FS, with dAuth enabling decentralized identities. This production launch validates J33VES as a robust alternative to proprietary AI assistants and positions Ratio1 to power similar enterprise deployments.

Deeploy v2 & Escrow v2 (R1‑348)

Building on Deeploy v1 (which provided full web UI/UX for container orchestration), v2 automates the entire job lifecycle. CSP users can create, edit and extend jobs via the UI, schedule start times and adjust parameters mid‑execution. The new job‑editing feature (R1‑443) eliminates the need to redeploy a container when tweaking configuration.

Deeploy now generates draft invoices for each completed job and provides burn reports showing how many R1 tokens were permanently burned as part of the job’s escrow settlement. This transparency aligns with PoAI’s on‑chain escrow system and ensures that compute costs and token supply adjustments are visible to all stakeholders.

The escrow smart contracts have been upgraded to fully automate rewards distribution to node operators and developers. Users deposit stablecoins into escrow; funds are allocated automatically once proof‑of‑AI verifies that the compute work is complete and burn reports are published.

LabelCraft v0.5 closed beta (R1‑125)

Crowdsourced data labeling: LabelCraft, part of the AI‑for‑Everyone (AI4E) initiative, gamifies data annotation and rewards volunteers with tokens. The v0.5 beta introduces improved tasks, enhanced quality‑control through multiple annotator consensus and a new leaderboard system. The app’s smart contract automatically distributes rewards once enough labels are collected and verified.

Integration with creation toolkit: LabelCraft seamlessly feeds annotations into the Creation Toolkit so that end‑users can build custom ML models on mobile devices via a guided wizard. The v0.5 beta also improves integration with decentralized training pipelines, enabling federated training over R1FS using homomorphic encryption.

Petware templates & cross native-generic applications (R1‑343)

Still in work Petware introduces a set of template applications that illustrate how Ratio1’s decentralized compute and storage can power both native and generic AI workflows. Templates include an end‑to‑end pipeline for building object‑recognition apps, a federated analytics dashboard and a low‑code starter kit for domain‑specific assistants based on computer vision. These templates, available soon, lower the barrier for developers to create cross‑platform AI apps and will be refined based on community feedback.

RedMesh v1 release (R1‑423)

Initially introduced last month, RedMesh is a decentralized, open‑source framework for continuous penetration testing. Jobs are announced across the network via CStore and executed by multiple edge nodes, allowing simultaneous scanning and vulnerability analysis. Findings are aggregated without a centralized server and results are stored on‑chain to ensure trust and immutability.

Unlike centralized commercial pentest platforms, RedMesh offers decentralized distributed execution, built‑in blockchain trust, open‑source transparency, continuous scanning, and resilience to single‑node failures. The release includes plugins based on various standards such as OWASP for service fingerprinting, web vulnerability scanning and compliance auditing, and forms the foundation for Ratio1’s cybersecurity offerings.

CerviSafe PoC (R1‑425)

CerviSafe, currently in proof‑of‑concept stage, is a framework for operationalizing decentralized cervical‑cancer screening. It uses a mobile‑first pipeline where clinics capture images locally; models are trained via federated learning with homomorphic encryption so that raw data stays on devices. Encrypted model updates are aggregated on the network and a global model is distributed back to clinics, providing accurate screening without exposing personal data.

The PoC responds to limited specialist availability and privacy concerns in underserved regions by enabling remote AI‑powered diagnosis. On‑chain traceability ensures that screening outcomes are transparent and verifiable while the EDIL framework (Encrypted Decentralized Inference and Learning) provides strong privacy guarantees.

Partnerships & Consortiums

Throughout September, Ratio1’s leadership participated in extensive meetings with international consortiums aiming to integrate decentralized AI into large‑scale projects. These consortiums include Pathfinder, European Innovation Council (EIC) and Horizon Europe initiatives focusing on healthcare, cybersecurity, energy and public‑sector innovation.

Prospective partnerships involve tens of companies across more than 15 countries, from academic institutions and research institutes to SMEs and multinational corporations. By collaborating with Ratio1, these partners plan to leverage decentralized compute (edge nodes), encrypted storage (R1FS), federated learning and token‑based incentives for fair collaboration.

Integrating Ratio1 into these consortiums can accelerate development of privacy‑preserving health diagnostics (e.g., expanded CerviSafe), continuous cybersecurity monitoring (leveraging RedMesh) and energy‑efficient AI services. Participation also positions Ratio1 as a key player in shaping EU‑level standards for trustless AI and fosters adoption across industries.

Outlook

September’s achievements lay the groundwork for an explosive final quarter. With an enterprise application live, core services upgraded and new proofs‑of‑concept demonstrating both cybersecurity and healthcare use‑cases, Ratio1 is transitioning from testing to real‑world adoption. In the coming months the team plans to:

Formalize agreements with Pathfinder, EIC and Horizon Europe consortia. Pilot joint projects in healthcare and cybersecurity to showcase Ratio1’s decentralization benefits.
Extend J33VES assistants to more customers and refine Keysoft’s production system. Explore integration with other enterprise platforms through standardized APIs.
Incorporate community feedback and launch Petware 1.0, enabling anyone to create cross‑platform AI apps with minimal code.
Continue improving the NodeJS clients and templates, release SDKs for other languages and provide more templates and tutorials.
Release new courses, intensify outreach on social media and host hackathons to onboard developers, researchers and non‑technical users.

By continuously delivering on roadmap goals and forging strategic partnerships, Ratio1 is positioning itself as the decentralized AI operating system for the next decade.