Live interactive graph
Cytoscape maps with pan, zoom, filters, multi-system switching, and service detail pages.
Onboard engineers in hours, not weeks of tribal knowledge.
Learn more →
Navo
Your architecture, alive.
YAML in Git becomes a living map and wiki. Graph, CLI, navo-mcp for agents. No platform team.
- Unlimited viewers free
- Git YAML as source of truth
- Static SPA, private-network ready
- MCP for AI agents
How Navo works
1
Describe
Write architecture in YAML — or generate a draft with navo scan / convert — and keep it in Git.
architecture.yaml
services:
- id: api-gateway
name: API Gateway
type: service
group: edge
links:
- auth-service
- user-service2
Compile
Navo validates schema and integrity, then builds a living architecture map for the SPA.
3
Explore
Explore the graph, open service detail, read wiki pages, annotate, and contribute via PR.
API Gateway
edge · service
OverviewLinksReposDocs
Central edge entrypoint routing auth and user traffic into the platform mesh.
Owner
Platform
Runtime
AKS
Repos
gateway-api
Wiki
ado.md
Capabilities
More than a diagram tool
Visualization, Git integration, AI imports, and agent-ready MCP - the living architecture platform for teams tired of stale slides and heavy IDPs.
Explore the Architecture Browser
Interactive graph, service detail, in-app wiki, multi-system maps, and Git export - one YAML source of truth.
Use cases
Built for platform teams, architects, and AI-era workflows
Whether you are escaping stale diagrams, avoiding Backstage fatigue, or feeding agents real topology - Navo keeps architecture living next to code.
Platform teams
Platform team at scale
Give every squad a shared living map of the platform without standing up Backstage or hiring a portal team. Git stays the source of truth; stakeholders browse for free.
- Onboard new engineers against a current map, not stale slides
- Review architecture changes in pull requests
- Share one URL for unlimited viewers behind VPN
Architects
Migration from Lucidchart / Confluence
Stop redrawing boxes every quarter. Import diagrams and generate YAML with AI, then keep the map alive next to the code that implements it.
- Brownfield systems become YAML maps faster
- Diagrams stop drifting from production reality
- Docs live in the in-app wiki beside each service
AI / platform leads
AI agents querying architecture
Moor Cursor, Claude, VS Code, and other MCP hosts to navo-mcp so agents list maps, walk graph paths, and validate drafts against the same Git YAML the SPA renders — not hallucinated topology.
- Typed tools for maps, nodes, edges, find_path, and validate
- Read-only by default; write and scan/convert opt-in
- navo:// resources and prompts for grounded agent workflows
Compare
Navo vs Backstage, IcePanel, and Lucidchart
Navo is the simplest Git-native architecture browser for teams that want a living map next to code - not a full internal developer portal and not another slide deck.
| Capability | Navo | Backstage | IcePanel | Lucidchart |
|---|---|---|---|---|
| Git-native source of truth | YAML in Git | Catalog + plugins | Cloud workspace | Cloud diagrams |
| Interactive living map | Partial (plugins) | Static / limited | ||
| In-app wiki next to the map | TechDocs | |||
| AI YAML / diagram imports | Varies | Limited | Import only | |
| navo-mcp for AI agents (tools + resources) | Emerging | |||
| CLI (bulk + AST → YAML) | Scaffolders | |||
| Unlimited free viewers | Self-host ops cost | Per-seat | Per-seat | |
| Deploy model | Static SPA / Docker | Heavy IDP stack | SaaS | SaaS |
| Time to value | Minutes | Months | Days | Hours |
| Requires platform team |
Want the deep dive? Navo vs Backstage
Built for teams who need the full picture
Architecture lives in slides and dies in Confluence.
Only two people know how the platform fits together.
Diagrams drift from the real system.
fremen-platform-ref · v1.0.021 services
Navo keeps your architecture living next to the code.
Git is the source of truth.
Validate in CI, deploy anywhere.
Static SPA - simple and secure.
Optional Elasticsearch for search.
Annotations stay local until you export for a PR.
Operator CLI for scan, convert, and deploy.
See Navo in action
Interactive maps in 2D and 3D, plus a CLI pipeline demo that keeps IaC, apps, and systems maps current via GitOps.
A workflow that fits how developers work
Edit YAML in Git
Define services, groups, edges, and docs — or seed maps with navo scan / convert.
Validate in CI
Catch issues early with schema and integrity checks.
Deploy Anywhere
Docker/nginx, operator CLI, Kubernetes, or static hosting.
Explore & Evolve
Graph, detail, wiki, annotate locally, and contribute via PR.
Built around Git, not around a platform team
Navo keeps architecture canonical in your repository. Search indexes and browser annotations are helpers - never a second source of truth.
Git remains canonical
Your architecture map lives in the repository. Navo reads compiled config - it never becomes a second source of truth.
Changes ship through pull requests
Annotate locally, export merged YAML, and open a PR. Reviewers validate schema in CI before merge.
Deploy as a static SPA
Ship with Docker and nginx, Kubernetes, or any static host. Private DNS and VPN-friendly for internal maps.
Unlimited viewers
Anyone with access can browse maps. Editors are the people who change architecture YAML - not every visitor.
FAQ
Frequently asked questions
Straight answers about Git-native architecture browsing, Backstage alternatives, and MCP.
What is a Git-native architecture browser?
A Git-native architecture browser treats architecture definitions stored in Git (usually YAML) as the system of record, then renders them as an interactive map teams can explore. Navo compiles YAML into a living Cytoscape graph with service details and an in-app wiki - without a separate SaaS diagram store.
How does Navo compare to Backstage?
Backstage is a full internal developer portal (catalog, scaffolder, plugins) that often needs a dedicated platform team. Navo is a focused Architecture Browser: YAML in Git → interactive map + wiki, with MCP for agents. Teams that just want the map use Navo in minutes instead of months of portal ops.
Does Navo support diagram imports and AST parsing?
Yes. Navo includes an AI YAML generator, diagram import paths, and CLI tooling for bulk imports and AST → YAML workflows so brownfield systems can become living maps without redrawing everything by hand.
What is Navo's MCP server for?
navo-mcp is Navo’s Model Context Protocol server. It exposes ArchitectureMap YAML to AI hosts through typed tools (list maps, load topology, find paths, validate drafts), navo:// resources, and workflow prompts. Agents get the same living map humans browse — with Git as the source of truth and read-only as the default.
Is Navo free for viewers?
Yes. Unlimited viewers can browse architecture maps on every plan. Community is $0 and includes the interactive graph, Git YAML workflow, wiki, navo-mcp, and AI YAML generator. Team meters editors only.
Does Navo require Elasticsearch or a cloud dependency?
No. Git YAML remains canonical. Elasticsearch is an optional secondary index. Deploy Navo as a static SPA with Docker/nginx or any static host - including private networks behind VPN or private DNS.
How does Navo keep architecture maps up to date automatically?
Through CI/CD pipeline integration. After every merge the Navo CLI runs navo scan on application and IaC trees, commits updated YAML back to Git, and lets Argo CD or Flux redeploy a locked-down Navo SPA in Kubernetes. Maps stay current without manual diagram editing.
Can Navo import diagrams from Lucidchart, draw.io, or Mermaid?
Yes. The operator CLI supports navo convert for Lucidchart JSON, draw.io XML, and Mermaid markdown. Each format is translated into Navo's ArchitectureMap YAML so existing diagrams become interactive, version-controlled maps immediately.
What is the Navo operator CLI?
The Navo operator CLI (navo) is a Go binary that automates architecture map workflows. Key commands include scan (AST-based codebase → YAML), convert (Lucidchart, draw.io, Mermaid → YAML), start/stop/status for local and Docker, and init for bootstrapping new projects. It runs in CI/CD pipelines or on developer machines.
Can I deploy Navo in a private Kubernetes cluster?
Yes. Navo is designed for private, locked-down deployments. Ship the static SPA as a Docker image with ClusterIP or internal ingress only, NetworkPolicy deny-by-default, non-root containers, and read-only root filesystem. No public cloud dependency or external SaaS is required.
How does Navo compare to IcePanel?
IcePanel is a SaaS-hosted C4 modelling workspace with per-seat pricing. Navo is a Git-native architecture browser you self-host: YAML stays in your repo, the interactive graph deploys as a static SPA or Docker image, and unlimited viewers browse for free. Navo adds an operator CLI for codebase scanning, MCP for AI agents, and an in-app wiki - without SaaS lock-in.
Does Navo work with AI agents and MCP?
Yes. navo-mcp works with any stdio MCP host — Cursor, VS Code, Claude Desktop, Claude Code, Gemini CLI, Grok, and similar agents. Default toolsets cover maps, graph queries (including find_path), and validation. Write and ops (scan/convert via the CLI) are opt-in behind explicit flags. Community includes MCP free on every plan.
What tools does navo-mcp provide?
Default read-only tools include get_server_info, get_workspace_context, list_maps, get_map_summary, get_map, search_maps, get_node, list_nodes, get_edges, find_path, list_groups, validate_map, and validate_yaml. Opt-in write_map, scan_codebase, and convert_diagram require --allow-write and toolset configuration. See /docs#mcp and /features/mcp.
Get involved
Unlimited viewers. Clear path from demo to Team.
Start with a live demo, read the docs, or talk to Fremen Labs about trials and enterprise deployments.
Keep architecture living next to code
Reduce architecture drift and meetings. Unlimited viewers free. Community includes MCP, AI YAML, and Git as source of truth.