J33VES x Keysoft x Ratio1 - Three Assistants That Turn Intent into Results

Privacy by default. Decentralized by design.

TL;DR

Keysoft has partnered with Ratio1 to launch three production-grade assistants on J33VES, Ratio1’s decentralized agent framework:

1. Description → Application

2. User need → Detailed report

3. Policy-grounded support chatbot

All assistants run on the Ratio1 network, with user-owned, end-to-end encrypted data stored in R1FS.

Why J33VES?

J33VES is Ratio1’s decentralized, trustless assistant framework. Its design principle is simple:

Your models and your data stay under your keys - not ours, and not anyone else’s.

Conversation histories, documents, and knowledge remain encrypted on the Ratio1 network, only accessible via the owner’s keys.

Under the Hood

• R1FS: An IPFS-style file system for private, encrypted storage and retrieval, paired with ChainStore (CStore) for decentralized state and discovery.

• dAuth: Decentralized authentication for nodes and pipelines.

• Open-weights LLMs & neuro-symbolic tools: Assistants connect to domain data and real workflows without central servers.

If you’re new to Ratio1, think of it as an AI operating system powered by blockchain - a kind of compute ride-sharing for AI.

The Keysoft Build: Three Assistants, One Framework

All three assistants share the same foundation:

• Zero trust by default: identities managed by dAuth, data encrypted in R1FS, auditability across the network.

• Bring-your-own schema/knowledge: documents, schemas, and policies live in R1FS, versioned and encrypted.

• Composable skills: J33VES pipelines layer tools (SQL formatter, parser, policy checker) around the model for reliable results.

1. From Description to Database (and UI)

The win:
Turn a plain-English idea into a clean, workload-aware database design - tables, indexes, constraints - that Keysoft can immediately render into a usable UI.

How it works:

• Drafts an entity-relationship plan.

• Applies naming and normalization heuristics.

• Outputs runnable MySQL DDL with indexes.

• Keysoft renders this into a ready-to-use UI/UX.

Example request:

“Multi-tenant subscription SaaS with invoices and usage metering”

Output snippet:

CREATE TABLE tenants (
  tenant_id INT PRIMARY KEY AUTO_INCREMENT,
  name VARCHAR(100) NOT NULL,
  email VARCHAR(254) NOT NULL,
  created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
  updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
);

CREATE TABLE subscriptions (
  subscription_id INT PRIMARY KEY AUTO_INCREMENT,
  tenant_id INT NOT NULL,
  plan_name VARCHAR(100) NOT NULL,
  start_date DATE NOT NULL,
  end_date DATE,
  status VARCHAR(50) DEFAULT 'Active',
  created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
  updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
    FOREIGN KEY (tenant_id) REFERENCES tenants(tenant_id)
);
-- …plus invoices, invoice_items, usage_metering

2. From Schema to Reports

The win:
Ask for insights (“Top customers by 2024 revenue”) and get safe, runnable SQL that only touches your known schema. Keysoft then generates the full report from your data.

How it works:

• Assistant ingests schema or introspection dump.

• SQL generation is constrained to known tables/columns.

• Queries are grounded in schema → far less hallucination than free-form generation.

Example:

Schema: customer(id, name), invoice(id, customer_id, total_cents, issued_at)
Request: “Top 10 customers by revenue in 2024”

SELECT c.id, c.name, SUM(i.total_cents) AS revenue_cents
FROM customer c
JOIN invoice  i ON i.customer_id = c.id
WHERE i.issued_at >= DATE '2024-01-01' AND i.issued_at < DATE '2025-01-01'
GROUP BY c.id, c.name
ORDER BY revenue_cents DESC
LIMIT 10;

3. Policy-Grounded Support Chatbot

The win:
Get clear, policy-compliant answers (with steps and references) every time - ideal for HR, finance, compliance, onboarding, and customer support. If no policy covers the request, the bot flags the gap instead of guessing.

How it works:

• Policies and knowledge (DOCX, TXT, PDF) live in R1FS, encrypted and versioned.

• Bot retrieves only what’s relevant to the question.

• Responses are strictly guided by retrieved rules/policies.

• Outputs are always text-only: instructions, checklists, references.

Example exchange:

• User: “How do I request a new laptop?”

• Agent: “Per Device Requests policy, employees and contractors can request a standard_laptop. Submit the form at /forms/device-request and open a ServiceDesk ticket. Your manager’s approval is required. Checklist: (1) fill the form, (2) attach justification, (3) notify your manager.”

Data Ownership & Privacy

• You own the keys. Conversations, schemas, and knowledge are encrypted under your keys; no central data lake.

• Peer-to-peer sync. Artifacts (briefs, schemas, policy packs, logs) are stored in R1FS, discovered via CStore. No single point of failure.

Why This Matters for Keysoft Customers

• Faster time to value: From idea → schema/UI. From question → report. From inquiry → policy-true answer.

• Lower risk, higher control: No central storage; encryption and continuous identity validation align with Zero Trust + E2EE.

• Scalable and portable: Content-addressed storage and decentralized identity make your assets easy to version, verify, and move.

Architecture at a Glance

1. J33VES Pipeline → LLM + tools (parser, SQL linter, policy checker).

2. Knowledge IO → R1FS for encrypted files; CStore for CID discovery.

3. Identity → dAuth for nodes and app-level plugins.

4. Execution → Ratio1 edge nodes coordinate workloads across the decentralized network.

Roadmap Impact

This deployment delivers two Ratio1 milestones:

• R1-6: Launch J33VES v1 (multi-foundation model decentralized framework) → Keysoft assistants run on J33VES v1, showing multi-model support and decentralized deployment.

• R1-26: Release customizable, white-label J33VES AI assistants on mainnet for enterprise → Keysoft’s branded assistants validate the white-label approach.

What’s Next

• RAG-as-a-Service on R1FS, powering grounded answers over large knowledge bases.

• Expanded SDKs: Node.js already supported for storage/state on Ratio1, with more coming.

In short: Keysoft and Ratio1 just showed how decentralized assistants can be production-ready, privacy-first, and enterprise-scalable.