ArchWiki

Turn any codebase into a visual architecture wiki for humans and agents. Scan repos, detect patterns, generate interactive HTML diagrams.

Why ArchWiki?

Understanding a codebase shouldn't require reading every file. ArchWiki scans your repo, detects routes, services, databases, auth, and build systems, then generates an interactive visual wiki showing how everything connects.

For humans — see the big picture before diving into code.

For agents — get structured architecture context for better code generation.

Install

npm install -g @talocode/archwiki

Quickstart

# Scan a repo
archwiki ./my-project

# Output:
#   .archwiki/architecture.json
#   .archwiki/architecture.html

Open architecture.html in your browser to see the interactive visual wiki.

CLI Commands

Scan repo

archwiki <repo-path> [output-dir]

Scans a codebase and generates architecture files.

Content-to-diagram

archwiki content <file.md> [output.html]

Convert long-form content (blog posts, docs, research papers) into visual diagrams. Extracts structure, identifies key ideas, and shows concept connections.

Supported formats: Markdown, plain text, HTML

What It Detects

| Pattern | Examples | |---------|----------| | Routes | Express, FastAPI, Flask endpoints | | Databases | SQL, MongoDB, Redis, Prisma | | Build | Docker, CI/CD, Makefile, pyproject.toml | | Auth | JWT, OAuth, API keys, middleware | | External | HTTP clients, gRPC, websockets | | Client | React, Vue, Angular, Next.js |

API Usage

from archwiki import ArchScanner, ContentDiagramScanner

# Scan a codebase
scanner = ArchScanner("/path/to/repo")
architecture = scanner.scan()

# Convert content to diagram
content_scanner = ContentDiagramScanner("# My Article\n\nContent here...")
diagram_data = content_scanner.generate_diagram_data()

Architecture Output

architecture.json

Structured data for agents:

{
  "name": "my-project",
  "nodes": [...],
  "edges": [...],
  "flows": [...],
  "stats": {
    "total_files": 42,
    "total_nodes": 12,
    "total_edges": 8,
    "total_flows": 3
  }
}

architecture.html

Interactive visual wiki with:

• Color-coded nodes by type
• Clickable elements with details
• Flow visualization
• Dark theme matching Talocode design

Content Diagrams

ArchWiki can also analyze long-form content:

1. Extract Structure — parses headings into hierarchical sections
2. Identify Key Ideas — scores sections by importance signals
3. Find Connections — detects cross-references and shared concepts

Output is an interactive HTML diagram showing how ideas connect.

Talocode Ecosystem

| Product | Install | Purpose | |---------|---------|---------| | ArchWiki | npm i -g @talocode/archwiki | (this repo) Codebase → visual wiki | | DocuLane | npm i -g @talocode/doculane | Office document parsing | | XSearchLane | npm i @talocode/xsearchlane | X/Twitter research | | CodeLane | npm i -g @talocode/codelane | Live coding streams | | FlowLane | npm i -g @talocode/flowlane | Visual workflow builder | | VideoLane | npm i -g @talocode/videolane | Video generation | | MailLane | npm i -g @talocode/maillane | Email API | | MemoryLane | npm i -g @talocode/memorylane | Agent memory | | Wiki | npm i -g @talocode/wiki | Knowledge base | | Tera | npm i -g @talocode/tera | AI chat API | | Codra | npm i -g @talocode-codra | Code review | | SearchLane | npm i -g @talocode/searchlane | Web search | | StackLane | npm i -g @talocode | Cloud platform |

Links

• GitHub: https://github.com/talocode/archwiki
• npm: https://www.npmjs.com/package/@talocode/archwiki
• Docs: https://docs.talocode.site

License

MIT © Talocode