StackRails indexes your code and its architecture across all your repos.
AI Coding Governance for the enterprise. StackRails gives every AI coding agent context from your code, architecture and policies, then enforces your policies with three checks: before the agent writes, on the developer's machine before code is committed or pushed, and at the pull request.
Before Code Is Written
The first check, before the agent writes. Every feature reaches your agents through one vendor-neutral MCP server (Claude Code, Copilot, Cursor, and Codex alike), giving the AI the wider context of your architecture and your organisation’s standards on every interaction. Context and policy go into the prompt so the first pass is steered toward compliant code.
Architecture-Aware AI Coding
Every agent generation steered by your architecture and standards.
Not generic training data. Every prompt to Claude Code, Cursor, Copilot, and Codex carries your architecture graph and the approved patterns for the product domain the agent is working in. Generations match your frameworks, your conventions, and your API style by default.
Cross-Service Context
The AI understands how your services connect.
Even across repositories. Agents correctly reference other services’ API contracts, events, and data models when they generate code, no hallucinated endpoints.
Codebase Context
See where your code fits in the wider system.
For any service or function: which domain it belongs to, what depends on it upstream and downstream, what gets impacted by changes, and which patterns apply. Visible in the Portal; available to your agents over MCP.
Explain This Code
Understand code in the context of your architecture.
Ask the AI to explain any code in terms of what it connects to and the architectural role it plays in the wider codebase.
Impact Analysis
See what depends on this code across repos.
Know what gets affected before you change it. The AI maps the full dependency chain across the entire codebase, across repositories.
Fix With AI
Ask your agent to make the code conformant.
When the conformance check surfaces a violation, tell Claude Code, Cursor, Copilot, or Codex to fix it. The agent gets the violation, the relevant ADR, and the approved pattern through MCP, then regenerates compliant code in place.
Generate From Architecture
Describe what you need. Get your patterns.
Get an implementation that uses your framework, your patterns, your conventions, your API style. Architecture-led generation, not generic scaffolding.
Migration Assistant
Move to new patterns across your codebase.
AI handles the implementation across repositories while you review. Migrate from deprecated patterns to approved ones at scale.
Conformance Report
Every code change scored against your standards.
Architecture, security, and standards compliance, all assessed automatically. Issues explained with links to the relevant ADRs and approved patterns.
Architecture Diff
See how a change affects the architecture.
Every structural change: new dependencies, boundary crossings, and pattern deviations, all surfaced explicitly so reviewers see the impact.
Auto-Review Comments
AI review comments referencing your standards.
Not generic rules. Every comment is contextual to the service and domain, referencing your specific architectural decisions and patterns.
CI/CD Gate
Block merges that violate critical rules.
Configurable thresholds per domain, team, or repository. Critical architectural violations never reach production.
On Every Code Change
The second and third checks. On the developer's machine, the AI's output is checked before code is committed or pushed; at the pull request, a conformance report and audit let the architect approve before merge. Developers get feedback before human review starts.
The Architecture Graph
The core. StackRails models how your enterprise is actually built (code and call graph, IaC, live cloud, CMDB, ADRs, and event-bus topology) across modern and legacy systems alike. Legacy systems you can’t easily change are ingested through their API endpoints even when there’s no source to read, and any legacy code you own is ingested too. The map stays current on every commit. This is the context that gives every agent a wider understanding of your systems.
Automatic Graph Construction
Point StackRails at your repositories. Done.
The Architecture Graph builds itself. No manual modelling required. Every service, function, dependency, and data flow is mapped automatically, across modern and legacy systems alike.
Incremental Updates
The graph stays current on every commit.
No batch rebuilds. Every push updates the graph incrementally, so the AI always works with the latest state of your system.
Graph Explorer
Navigate your architecture visually.
Services as nodes, dependencies as edges, data flows as paths. Filter by domain, team, technology, or conformance status.
Architecture Queries
Ask questions about your system in natural language.
“What services write to the customer database?” “What would break if we change this schema?” Get answers instantly from the Architecture Graph.
Auto-Generated Documentation
Architecture docs from the code. Always current.
Generated from the actual codebase, not from memory. Never stale. Always reflects what’s really running in production.
Pattern Detection
Identifies patterns your codebase already follows.
StackRails surfaces the patterns in your code. You review and promote the ones that should be standards. Data-driven, not opinion-driven.
Product & Domain Scoping
Different products get different standards.
The AI knows which domain a developer is working in and applies the right rules. Your payments platform has different patterns from your IoT edge services.
Reference Repositories & Code Curation
Mark your best repos as the gold standard.
Flag legacy code. The AI learns from the code you want to replicate, not from the code you’re migrating away from.
Pattern Confidence Scoring
Data-driven standards adoption.
See how consistently a pattern is followed, where it comes from, and whether competing approaches exist. Make informed decisions about what to standardise.
Governance & Policy
The reason to buy. Your standards become an executable Policy Pack, not PDFs nobody reads, enforced at all three checks. Most enterprises start with repo-level markdown rules and prompt files; StackRails is the enterprise version: governed, versioned, permissioned, and enforced. Your architecture team sets the rules; StackRails makes them self-enforcing across every tool, team, and repo.
Policy Editor
Define architectural rules per domain or org-wide.
Enforced by all three checks: before the agent writes, on the developer's machine, and at the pull request. Rules are versioned, auditable, and easy to update.
Pattern Library
Your approved patterns, organised and versioned.
Curated from reference code, organised by domain, browsable by any developer. The single source of truth for how things should be built.
ADR Tracking
Know which decisions are being followed.
Track which architectural decisions are being followed and which are drifting. Connect decisions to the code that implements them.
Policy Inheritance & Exceptions
Org baselines with team-level overrides.
Org-wide baselines, product-level standards, team-level overrides, all auditable, all with expiry on exceptions. No permanent workarounds.
Technology Radar
Track approved, tolerated, and deprecated tech.
See which technologies are approved, tolerated, or deprecated across your codebase. Enforce technology choices through the AI.
Dashboard & Reporting
Proof that it’s working. Visibility for leaders, and evidence you can hand to your own auditors. (StackRails produces evidence for your audits; it is not itself a certification.)
Conformance Dashboard
Architectural health across the organisation.
Drill from overview to individual code changes. See conformance trends, identify improving and declining teams, and track standards adoption.
Team Comparison
Compare conformance across teams.
Identify which teams are leading in architectural conformance and which need support. Data-driven engineering management.
Onboarding Metrics
Measure time to architectural conformance.
Track how fast new developers reach architectural conformance. Prove that StackRails accelerates onboarding.
Risk Hotspots
See where architectural debt concentrates.
Identify the areas of your codebase with the highest architectural risk. Prioritise remediation based on data, not guesswork.
Audit Evidence Export
Evidence for your own audits, from your codebase.
Generate evidence to support your own SOC 2, ISO 27001, and internal audits, backed by data from the actual codebase. StackRails helps you prepare for your audits; it does not certify you and is not a certification itself.
Executive Summary Reports
Monthly quality reports that write themselves.
Engineering quality reports generated automatically. Share with leadership without manual data gathering.
Enterprise Platform
Gated by your permissions and never used to train a model: the security, deployment, and integration requirements your IT and procurement teams need checked.
SSO & Provisioning
SAML, OIDC, SCIM
Role-Based Access Control
Granular permissions
Audit Log
SIEM export
Deployment Options
Hosted EU / VPC / on-prem / air-gapped
Encryption
AES-256 at rest, TLS 1.3 in transit
No Training on Your Code
Your code is never used for model training
SCM Integrations
GitHub, GitLab, Bitbucket, Azure DevOps
CI/CD Integrations
GitHub Actions, GitLab CI, Jenkins, Azure Pipelines
API-First
REST & GraphQL
SDK for Custom Rules
Extend with your own logic
Questions about the features
What is the Architecture Graph?
It is a model of how your systems are actually built: code and call graph, infrastructure as code, live cloud, CMDB, ADRs and event-bus topology, across modern and legacy systems. It builds itself from your repositories with no manual modelling, and it stays current on every commit.
How does StackRails do dependency mapping across repositories?
Every service, function, dependency and data flow is mapped automatically. For any service or function you can see what depends on it upstream and downstream, and what a change would affect across repositories, including loosely-coupled event-bus subscribers.
Does StackRails generate software architecture documentation?
Yes. Architecture documentation is generated from the code itself: service maps, ownership, dependency chains and API contracts. It is current on every commit, because it comes from what is actually running.
How do Claude Code, Cursor, Copilot and Codex get this context?
Through one vendor-neutral MCP server. The agent gets context from your code, architecture and policies, plus the approved patterns for the product domain it is working in. What it generates then matches your frameworks, your conventions and your API style.
What is a Policy Pack?
Your approved patterns, coding conventions, security policies and technology choices, written as executable rules. Rules are set per domain or org-wide, are versioned and auditable, and are enforced by three checks: before the agent writes, on the developer's machine, and at the pull request.
Can StackRails block a pull request that breaks our rules?
Yes. Every code change is scored against your standards in a conformance report, with each issue linked to the relevant ADR and approved pattern. A CI/CD gate blocks merges that violate critical rules, with thresholds you set per domain, team or repository.