Back to Overview
Solveaux Platform • Organizational Intelligence

Give your organization
a permanent memory.

Solveaux is the dedicated organizational intelligence engine for modern engineering teams. Capture what was decided, why it happened, and what came next — across people, projects, and autonomous AI agents.

The Architectural Memory Gap

Engineering decisions are made in fleeting meetings, ephemeral Slack threads, and closed pull requests. Months later, when systems scale or engineers transition, nobody recalls why specific trade-offs were chosen, resulting in repeated debates, architectural drift, and accidental regressions.

The Solveaux Solution

A centralized intelligence hub uniting Architectural Decision Records (ADRs), interactive dependency graphs, Git-native .slvx blocks, and direct Model Context Protocol (MCP) support for AI coding assistants.

Inside Every Project: The 8 Dedicated Intelligence Tabs

Each project workspace equips engineering teams with a unified operational suite to capture, validate, and execute architecture.

TAB 01

In Progress

Sprint Kanban

Active sprint execution linked directly to architecture.

Manage tasks with priorities (Urgent, High, Medium, Low) across standard workflow states (Todo, In Progress, In Review, Done). Every sprint item can link directly to corresponding ADRs, ensuring developers build code that honors established system constraints.

Sprint Kanban board with live priority & assignee filters
Direct bi-directional links to Architectural Decision Records (ADRs)
Prevents architectural divergence during fast-paced sprints
TAB 02

Research

Technical Spikes

Benchmark investigations & hypothesis evaluations.

Document technical spikes, library bake-offs, latency benchmarks, and architectural investigations before committing code. Structured notes capture problem statements, hypotheses, candidate options, and empirical findings.

Benchmark test logging with latency and throughput metrics
Structured technical notes with code snippets and findings
Preserves spike conclusions so teams never re-run the same experiment
TAB 03

Decisions (ADRs)

Decision Records

The immutable source of truth for architectural choices.

Permanent Architectural Decision Records (ADRs) capturing context, status (Proposed, Accepted, Superseded, Rejected), deciders, accepted trade-offs, and explicitly recorded rejected options. Eliminates circular team debates and organizational amnesia.

Structured status lifecycle with decider attribution
Captures rejected alternatives to stop repeated mistakes
Versioned history with clean superseding workflows
TAB 04

Validation

System Health

Empirical verification against architectural drift.

Continuous validation tests active production performance against past architectural assumptions. Formulate clear hypotheses, define target metrics (e.g. sub-50ms latency), and log empirical outcomes before releasing critical services.

Validation statuses: Verified, Pending Review, or Deprecated
Prevents silent tech debt accumulation across multi-year lifecycles
Flags outdated architectural assumptions before production outages
TAB 05

Feedback Hub

User & Team Insights

Customer & internal feedback converted into action.

A dedicated repository for user testing feedback, customer observations, and internal developer notes. Extract key quotes and convert actionable feedback directly into sprint tasks or ADR proposals with one click.

Quote highlighter to extract decisive customer sentiment
Convert raw feedback directly into Kanban tasks or ADRs
Centralizes user friction points alongside technical architecture
TAB 06

Brain Map

Knowledge Graph

Interactive visual node network of your entire system.

An interactive 2D/3D visual knowledge graph dynamically compiling tasks, research notes, ADRs, validation experiments, and services into an interconnected graph. Reveals dependencies and provides instant blast-radius impact analysis.

Visual node network mapping cross-service relationships
Instant impact analysis before refactoring or deprecating systems
Zero-context-switching onboarding for new engineers and tech leads
TAB 07

Team & RBAC

Access & Governance

Granular permissions & contributor accountability.

Manage project contributors with granular Admin, Author, and Viewer permissions. View author attribution on every decision record, assign tasks to engineers, and preserve complete audit trails of who made what architectural choice.

Granular role-based governance: Admin, Author, and Viewer
Soft-delete safeguards protect authorship history when engineers transition
Cryptographic single-use invites and team roster controls
TAB 08

Settings & MCP

Integrations & Sync

Model Context Protocol keys & Git-native .slvx sync.

Configure project metadata, generate API keys for Model Context Protocol (MCP) clients (Cursor, Claude Desktop, Windsurf), and manage bi-directional synchronization with Git-native .slvx block files in your code repository.

Project-level API key management for autonomous AI agents
Bi-directional synchronization with local .slvx Git directory
Project deletion safeguards and configuration controls

Organizational Intelligence Architecture

A comprehensive suite engineered to capture, connect, and operationalize engineering memory.

Core Foundation

Architectural Decision Records (ADRs)

Permanent, structured decision records capturing rationale, context, deciders, accepted trade-offs, and explicitly rejected alternatives. Prevent teams from re-debating settled architectural choices.

Structured status lifecycle: Proposed, Accepted, Superseded, Rejected
Captures rejected alternatives with reasons to stop repeated mistakes
Links directly to corresponding sprint tasks and validation experiments
System Intelligence

Brain Map & Interactive Knowledge Graph

An interactive visual node network connecting multi-project workspaces, microservices, ADRs, technical research, and team ownership into an interconnected system graph.

Visual dependency tracking across repositories and services
Instant blast-radius and impact analysis prior to refactoring
Zero-context-switching architectural exploration for tech leads
AI Integration

AI Agent Memory via MCP Protocol

Model Context Protocol (MCP) server bringing your organization’s active architectural constraints directly into Cursor, Claude Desktop, and Windsurf.

AI agents query accepted ADR rules before generating code
Automatic guardrail checks prevent hallucinated anti-patterns
Contributor agents propose ADRs held safely for human approval
Open Spec

Git-Native .slvx Specification

Human-readable, block-based Markdown format stored directly in your Git repository under .slvx/. Machine-parseable with zero cloud token consumption and seamless pull-request reviews.

Version-controlled with your code; zero proprietary vendor lock-in
Bi-directional synchronization between local repo and Solveaux
Deterministic state-machine parsing with zero cloud AI bills
Multi-Tenancy

Multi-Project Organization Workspaces

Consolidate every client application, backend service, microservice, or SDK under a unified organization umbrella with isolated project workspaces and shared intelligence.

Isolated research logs, ADR vaults, and task kanbans per project
Cross-project knowledge search and shared architectural standards
Granular role-based access control (Admin, Author, Viewer)
System Health

Continuous Validation & Drift Prevention

Architectural choices that work at early stages break under scale. Solveaux enables continuous validation cycles that test production reality against historical assumptions.

Systematic status tracking: Verified, Pending Review, or Deprecated
Clean superseding workflows when upgrading architectural foundations
Preserves technical institutional memory across multi-year lifecycles
Connected Workflow

How Decisions Flow Through Your Organization

From the initial technical spike to automated verification and AI context injection, every step is linked.

AcceptedADR-014
PostgreSQL RLS Multi-Tenancy

Context: Multiple projects and team members require strict data isolation without error-prone manual application filters.

Decision: Enforce database-level PostgreSQL Row-Level Security (RLS) linked to organization membership tokens.

Consequences: Zero risk of cross-tenant data leaks; query plans automatically scoped by database engine.

src/lib/supabase.tsmigrations/003.sql
Solveaux MCP Server
Claude / Cursor

// AI query before modifying database query

> get_file_context("src/lib/supabase.ts")

Active Constraint: ADR-014 Enforced
"Always pass authenticated user JWT. Do not bypass RLS using service role credentials in client-facing endpoints."

✓ AI generates compliant code without violating company architectural standards.

The 4-Stage Organizational Intelligence Cycle

A repeatable discipline ensuring institutional memory compounds with every sprint.

STAGE 01

Technical Spikes & Research

Validate hypotheses before committing code.

Conduct exploratory benchmarks, profile candidate frameworks, and document technical trade-offs with structured metrics. Ensure engineering choices are backed by data rather than intuition.

STAGE 02

Codify the Decision (ADRs)

Lock in rationale and document rejected options.

Formulate Architectural Decision Records with explicit deciders, status, and trade-offs. Document discarded alternatives so the team never repeats circular debates over past considerations.

STAGE 03

Align AI Agents & Codebase

Feed architectural memory into daily workflows.

Export decisions to Git-native .slvx blocks and expose them to Cursor, Claude, and Windsurf via the Solveaux MCP server. AI agents generate code that adheres to your established patterns.

STAGE 04

Continuous Validation & Knowledge Graph

Keep institutional knowledge fresh and connected.

Visualize all decisions on the interactive Brain Map. Continuously test production systems against past architectural assumptions to catch architectural drift before it creates technical debt.

Enterprise Privacy, Tenant Isolation & Data Sovereignty

Your architectural decisions, proprietary code paths, and benchmarks are strictly confidential.

Database-Level Isolation

PostgreSQL Row-Level Security ensures that organization records can never leak across tenant boundaries.

Role-Based Governance

Granular Admin, Author, and Viewer permissions regulate who can approve ADRs, export tokens, or invite team members.

Zero Public Exposure

Workspaces are private by default. Search crawlers and unauthenticated viewers cannot view your architectural intelligence.

Our Guiding Principles

The philosophical commitments that drive Solveaux’s engineering and product choices.

01

Knowledge Belongs to the Organization

When key engineers leave or teams reorganize, architectural reasoning, trade-offs, and critical context must remain intact. Solveaux makes technical intelligence institutional, permanent, and accessible.

02

Structure Over Ephemeral Noise

Unstructured Slack channels and pull-request comments generate noise, not knowledge. Structured decision records and linked dependency graphs retain their utility for years to come.

03

AI Must Respect Architectural Intent

Generating code at 10x speed without architectural context creates unmanageable technical debt. Solveaux equips autonomous AI coding agents with the exact rules and historical ADRs of your organization.

04

Living Graph Instead of Stale Docs

Static documentation graveyards gather dust and become obsolete. Connected knowledge graphs dynamically reflect projects, tasks, research, and validation health as your architecture evolves.

Enterprise Early Access

Stop losing what your engineering team knows.

Create an organization workspace, invite your team, connect your codebase via Git or MCP, and start building a permanent architectural memory today.