Your AI tools, orchestrated.
Lex turns any AI coding assistant into a coordinated team of specialists. Every stage of the SDLC, governed and handled.
Built by engineers who shipped platforms for
Capabilities
Not another AI coding tool.
A coordinated development team.
Lex transforms your existing AI tools into an orchestrated system that understands your project, your standards, and your workflows.
The Conductor Pattern
Lex never writes code. It routes requests to specialists, tracks multi-step workflows, chains outputs, and enforces quality gates.
Specialisation Over Generalisation
Org-defined agents, each expert in one domain. Design extraction, architecture, planning, implementation, review, and infrastructure -- customisable to your needs.
Declarative Workflows
YAML-defined pipelines that declare steps, agents, failure handling, and human checkpoints. The sheet music Lex follows.
Quality Gates Built In
Every change passes through validation. TypeScript, ESLint, tests, architecture review, security, accessibility -- all automated.
Living Knowledge Base
Shared instruction files define your standards. Update one file, every agent adapts. One source of truth, zero drift.
Self-Correcting Loops
When review finds issues, agents fix them automatically. TypeScript errors, lint violations, and architectural concerns are routed back with feedback.
The Pipeline
From Figma to production
in one sentence.
The design-to-production workflow: five steps, five specialists, zero context lost.
Extract Design Specs
Connects to Figma via MCP. Maps every colour, spacing, and font to your Tailwind tokens. Identifies interactive states, responsive breakpoints, and accessibility requirements.
maintainer-design
Plan the Work
Synthesises design specs into a PRD, execution plan, and Jira tickets with complexity estimates. The plan feeds forward into implementation.
maintainer-analysis
Build Everything
Researches the codebase, mirrors existing patterns, builds types, component, stories, and tests. Runs TypeScript and ESLint validation. Self-corrects errors.
executor-code
Quality Gate
Final validation across 10 areas: static analysis, architecture, error handling, security, design tokens, accessibility, performance, tests, stories, and feature flags.
executor-review
Document & Index
Creates component-level documentation, updates the symbol index so the new component is discoverable by other agents and engineers.
maintainer-architecture
Total time: ~9 minutes
Manual equivalent: 2-4 hours of context-switching
Agent System
Your agents. Your rules.
Lex doesn't ship a fixed roster. Your organization defines which agents exist, what they can do, and how they collaborate. Start with battle-tested defaults or build your own from scratch.
Design Interpreter
Connects to Figma via MCP, extracts component hierarchy, maps tokens, and identifies states and breakpoints.
Architecture Oracle
Traces code paths, creates ADRs, maintains institutional memory through the symbol index.
Code Builder
Builds in strict order: types, component, stories, tests. Mirrors patterns. Self-corrects on failure.
Review Gate
Final quality gate across validation areas. Has edit authority for fixes. Routes issues back to executors.
SRE Operator
Owns Terraform, CI/CD, Docker, and cloud. Security-first with least-privilege IAM and full promotion flow.
Your Custom Agent
Define any specialist your team needs. Set its rules, scope, tools, and review gates. Lex orchestrates the rest.
How it works
Fully dynamic, fully governed
The agent roster is configuration, not code. Add a new specialist in minutes and Lex handles orchestration, routing, and audit logging automatically.
Read the docsOrg-defined agent roster
Create, modify, and retire agents as your needs evolve.
Role-scoped permissions
Each agent only accesses the repos, tools, and workflows you allow.
Structured I/O contracts
Every agent reads the same rules format and returns structured, auditable results.
Composable specializations
Combine Maintainer (research, strategy) and Executor (build, ship) roles freely.
Workflow Catalogue
Just say what you need.
Lex picks the right workflow.
15 declarative workflows covering the full development lifecycle. Intent matching routes your request automatically.
Full Pipelines
design-to-productionrefactorStandard Flows
bug-fixperformanceFocused Tasks
quick-frontendtest-onlycode-reviewstorybookinfrastructurecodebase-healthSingle Expert
documentationadr-researchagent-configBuilt for Your Scale
Everything adapts to your organisation.
No fixed limits. Agents, workflows, integrations, and tests are all defined and controlled by your team.
Specialised Agents
Custom-built by your org. Define roles, scopes, and permissions to match your engineering structure.
Declarative Workflows
From single-expert tasks to multi-step orchestrated pipelines -- configured to your standards.
Integration Tests
Automatically generated and enforced to prevent config drift and routing regressions.
MCP Integrations
Connect the tools your team already uses -- Figma, Jira, Chrome DevTools, Playwright, and more.
Developer Experience Impact
The value isn't that AI writes code faster. The value is that AI writes code that belongs -- code that follows your conventions, passes your quality gates, and is documented and tested from day one.
Plans
Governed AI at every stage of growth
Every tier builds on the last. Choose the level of governance, architecture enforcement, and control your organization needs.
Standard
Governed AI for a Single Engineering Team
Deploy AI across your engineering team without losing control.
Core Platform
- Lex orchestration engine
- Declarative workflow system
- Specialised agents (planner, coder, reviewer, architect)
- VS Code extension + CLI
Governance
- Centralized workflow enforcement
- Mandatory review gates on all mutating workflows
- Activity logging with per-session trace
- Organization-level access controls
Model Strategy
- BYOK — Bring Your Own Key (Claude, GPT, Gemini, Ollama)
- Per-workflow model selection
- Secure credential isolation per provider
Observability
- Usage metrics per workflow & agent
- Adoption tracking across team
- Basic cost visibility per model
Guarantees
- No AI-generated code reaches production unreviewed
- Every AI action is logged and traceable
- AI follows your documented engineering standards
Governance & Architecture
Standardized Engineering at Scale
AI accelerates delivery while enforcing architecture, standards, and policy across the organization.
Everything in Standard, plus:
Advanced Governance
- Runtime policy engine with configurable rules
- Human-in-the-loop approval gates
- Violation detection & alerting
- Full audit trail with action-level detail
Architecture Intelligence
- Architecture agent with ADR generation
- Dependency mapping & impact analysis
- Refactor safety analysis
- Cross-repository pattern enforcement
Control Plane
- Executive visibility dashboard
- Risk & compliance reporting
- Team & workflow analytics
- Policy configuration UI
Advanced Model Strategy
- Policy-based model routing
- Sensitivity-aware model restrictions
- Cost guardrails per team & workflow
Guarantees
- Architecture decisions are documented and enforced
- AI cannot bypass security or engineering policy
- Leadership has full visibility into AI-driven change
- AI usage is defensible in audits and postmortems
Enterprise Control Plane
Mission-Critical AI Infrastructure
AI is a governed, auditable, cost-controlled part of your SDLC across the entire organization.
Everything above, plus:
Enterprise Governance
- Org-wide policy inheritance & cascading rules
- Environment-specific controls (dev / staging / prod)
- Release governance workflows
- Compliance-ready audit exports (SOC 2, ISO 27001)
Cost & Spend Intelligence
- Cost breakdown by workflow, repository & agent
- Optimization recommendations
- Budget alerts & hard caps
- Historical spend analysis & trends
Security & Deployment
- SSO via OIDC + SAML 2.0 (Okta, Azure AD, Google Workspace)
- SCIM user & group provisioning
- Encrypted prompt & output storage at rest
- Dedicated tenant isolation
Support & Reliability
- Incident response & rollback workflows
- SLA-backed support
- Dedicated customer success manager
- Private deployment option (VPC / on-prem)
Guarantees
- AI usage is compliant, auditable, and predictable
- Leadership can explain every AI-driven change
- AI spend is controlled, not reactive
- Lex becomes essential engineering infrastructure
Seat Types
Developer Seat
- Executes workflows
- Uses agents
- Writes code with Lex
Power Seat
Tech Lead / Staff / Manager
- Approves workflows
- Overrides gates
- Views advanced insights
- Configures policies
Get Started
Give your team
an AI development team.
Lex layers on top of your existing AI tools. No migration, no lock-in. Just better-orchestrated AI that understands your codebase.
Works with any AI coding tool. VS Code extension. Configuration-driven.