# FLOWCONSOLE — Full Product Description

> Architecture as Code platform. Model your software systems in code, visualize them automatically, detect drift between model and infrastructure, run fitness functions in CI/CD, and query the architectural graph with AI.

Contact: v@flowconsole.tech | GitHub: https://github.com/flowconsole | Twitter: https://twitter.com/vmgolubev

---

## What Is FLOWCONSOLE

FLOWCONSOLE is an Architecture as Code platform that transforms the way engineering teams document and validate software architecture. Instead of maintaining diagrams manually in graphical editors, teams describe their systems in familiar programming languages. The model is stored in Git, versioned, and automatically kept in sync with the actual codebase and infrastructure.

The core idea: the architectural model is a **queryable live graph**, not a collection of pictures. Diagrams are a byproduct of the model, generated automatically. This means the architecture is always up to date, machine-readable, and validatable.

## Core Capabilities

### Code-First Modeling (Architecture as Code)
Describe system elements (services, databases, APIs, teams) and their relationships using the `.fc.ts` DSL in TypeScript, or equivalent SDKs for C#, Go, Java, Python, or YAML. The model is a regular code file stored in Git.

### Automatic Diagram Generation
From the model, FLOWCONSOLE generates interactive C4-level diagrams (Context, Container, Component, Code). No manual drawing — the diagram is always consistent with the code.

### Drift Detection
FLOWCONSOLE compares the target architectural model (to-be) with the actual state of infrastructure and code (as-is). Discrepancies between them are matched element by element and ranked by impact. Supports Kubernetes, OpenAPI, Docker Compose, and Terraform as infrastructure sources.

### Automatic Code Scanning
FLOWCONSOLE scans source code to automatically extract architectural elements: services, APIs, data stores, dependencies. Currently supports C#; TypeScript, Go, Java, and Python are in active development.

### CI/CD Fitness Functions
Define architectural rules (fitness functions) that run in CI/CD pipelines. Examples: "no service should have more than 5 direct dependencies", "all public APIs must have an ADR". Violations fail the build.

### Living Architecture Model
The model is automatically updated via Git sync. Every commit to the architecture model file triggers a graph rebuild. The model always reflects the team's current decisions.

### AI-Powered Queries
Ask questions about the architecture in natural language. FLOWCONSOLE uses a locally hosted LLM to translate natural language queries into graph queries and return structured answers — no data leaves your infrastructure.

### Architecture Decision Records (ADR)
Create and link ADRs to specific architectural elements using the MADR format. ADR violations are detected automatically when the architecture diverges from the decision.

### Import from External Sources
Import existing architecture knowledge from Structurizr DSL, Backstage catalog-info.yaml, or CMDB CSV files. Merge or replace the existing model.

### MCP Server
FLOWCONSOLE exposes a Model Context Protocol (MCP) server at `/mcp`, enabling AI assistants (Claude Desktop, Cursor, VS Code Copilot) to query the architectural model directly.

---

## Supported Languages for DSL

| Language | DSL Format | Package |
|----------|-----------|---------|
| TypeScript | `.fc.ts` | `@flowconsole/sdk` |
| C# | `.fc.cs` | NuGet (FLOWCONSOLE.Sdk) |
| Go | `.fc.go` | Go module |
| Java | `.fc.java` | Maven/Gradle |
| Python | `.fc.py` | PyPI |
| YAML | `.fc.yaml` | built-in |

---

## Use Cases by Persona

**CTO / Engineering Director**
- Get a living architecture map of the entire system without asking engineers to draw it
- Track architectural drift and technical debt automatically
- Ensure architecture decisions (ADRs) are actually implemented

**Software Architect**
- Define and enforce architectural fitness functions in CI/CD
- Query the architectural graph: "which services depend on the payment service?"
- Generate C4 diagrams for stakeholder presentations automatically

**Tech Lead**
- Scan existing codebase to bootstrap an initial architectural model
- Detect when infrastructure diverges from the planned architecture
- Link code changes to architectural decisions

**DevOps / Platform Engineer**
- Scan Kubernetes clusters to populate the infrastructure layer of the model
- Validate service mesh topology against the architectural model
- Trigger drift detection on every deployment

---

## Comparison with Alternatives

| Feature | FLOWCONSOLE | Draw.io | Structurizr | IcePanel | LikeC4 |
|---------|------------|---------|-------------|----------|--------|
| Code-First Modeling | Yes | No | Yes (DSL only) | No | Yes |
| Interactive Diagrams | Yes | No | Partial | Yes | Yes |
| Git-Native Workflow | Yes | Partial | Yes | No | Yes |
| Drift Detection | Yes | No | No | No | No |
| CI/CD Fitness Tests | Yes | No | Partial | No | No |
| Multi-Language DSL | Yes (6 langs) | No | No (Structurizr DSL only) | No | No |
| Automatic Code Scanning | Yes | No | No | No | No |
| Living Architecture Model | Yes | No | No | Partial | No |
| CMDB/Backstage Import | Yes | No | No | No | No |
| Self-Hosted | Yes (Apache 2.0) | Yes | Yes | No | Yes |
| AI Queries | Yes (local LLM) | No | No | No | No |

---

## FAQ

**What is Architecture as Code?**
An approach where the architectural model of a system is described in programming code using familiar languages rather than drawn manually in graphical editors. The model is stored in Git, versioned, and available for analytical queries — as a living dependency graph.

**How does drift detection work?**
FLOWCONSOLE compares the target architectural model (to-be) with the actual state of infrastructure and code (as-is). Discrepancies between them are matched element by element and ranked by impact, so you can see what changed, where, and how serious it is.

**Is FLOWCONSOLE open source?**
Yes. FLOWCONSOLE is moving to open source under the Apache 2.0 license: the SDK will be released first, followed by the engine.

**How do I get started?**
Contact support to request an account. Once it's set up, describe your architecture model directly in your Git repository — FLOWCONSOLE will pick it up and visualize it. A detailed guide is in the documentation.

**What AI model is used?**
Only self-hosted models, no data leaves our infrastructure.

---

## Contact & Links

- Website: https://flowconsole.tech
- Documentation: https://flowconsole.tech/en/docs/
- GitHub: https://github.com/flowconsole
- Email: v@flowconsole.tech
- Twitter/X: https://twitter.com/vmgolubevv
- Founder: Vladimir Golubev, Founder & CEO