OpenTalk2HTML-NotMD MCP Server

Generate, patch, read, and compress HTML pages for reports, dashboards, and docs with fast AI-agent speed and precise CSS selector updates.

AI talks to you through beautiful pages — not overwhelming markdown dumps.

OpenTalk2HTML-NotMD is an MCP server that transforms how AI agents communicate with you. Instead of drowning in raw markdown files that nobody can read, AI builds clean, organized HTML pages — reports, dashboards, landing pages, invoices and more — delivered instantly, no browser required.

One command. Any AI platform. Real pages.

text

npx -y @aimino/opentalk2html-notmd

---

When to Use

• Generating complex HTML pages (reports, dashboards, landing pages, docs) from structured component specs
• Patching existing HTML by element id or CSS selector without re-rendering the entire page
• Compressing HTML to fit within token limits while preserving semantic content
• Maintaining cross-section consistency across large multi-section documents
• Building equity research, financial summaries, or data-heavy pages with charts and tables
• Creating email newsletters, pitch decks, changelogs, or API documentation

---

Quick Start

bash

npx -y @aimino/opentalk2html-notmd

Add to MCP client config:

Claude Desktop / Cursor / VS Code Copilot

json

{
  "mcpServers": {
    "open-talk-2-html-not-md": {
      "command": "npx",
      "args": ["-y", "@aimino/opentalk2html-notmd"]
    }
  }
}

Claude Code

bash

claude mcp add open-talk-2-html-not-md -e npx -a "-y" -a "@aimino/opentalk2html-notmd"

---

Workflow

1. Discover what's available

text

list_templates
list_components
get_template_schema(template: "report")
get_component_schema(component: "hero")

2. Build a page

text

render_page(
  template: "report",
  sections: [
    {component: "hero", props: {title: "Q3 Report", badge: "Draft"}},
    {component: "data-table", props: {headers: ["Metric","Value"], rows: [["Revenue","$1.2M"]]}}
  ],
  output_path: "/tmp/report.html",
  options: {title: "Q3 Report"}
)

3. Inspect the output

text

read_html(path: "/tmp/report.html", mode: "compressed")

Read modes (by token efficiency):

• structure — tree overview (70% savings)
• content — typed blocks (74% savings)
• compressed — summary + stats (87% savings)
• text — plain text only (97% savings, best for token budgets)

4. Refine with surgical patches

text

patch_html(file_path: "/tmp/report.html", selector: "#content", html: "<p>Updated content</p>")
set_attribute(file_path: "/tmp/report.html", selector: "#main-title", attr: "class", value: "highlight")

5. Compress for token efficiency

text

read_html(path: "/tmp/report.html", mode: "compressed", offset: 0, limit: 1000)

6. Maintain consistency (multi-section documents)

text

check_consistency(path: "/tmp/report.html")
propagate_edit(path: "/tmp/report.html", entity: "Revenue", value: "$1.5M", sections: ["summary", "details"])

---

Five-Tier Architecture

text

OpenTalk2HTML-NotMD MCP Server
├── Assembly   → render_page (compose pages from components)
├── Patch      → patch_html, set_attribute, edit_html_range (surgical edits)
├── Read       → read_html (inspect with 97% token savings)
├── Raw        → write_raw_html, format_html, preview_html
├── Email      → render_email, html_to_email (send to inbox)
└── Consistency→ propagate_edit, check_consistency (keep data in sync)

18 tools, 22 components, 25+ templates — purpose-built for AI-driven page creation with sub-second patch times and AI-grade token compression.

---

Resources

• references/ — detailed examples and advanced use cases
• scripts/ — automation scripts
• assets/ — templates and resources
• GitHub: https://github.com/Aimino-Tech/OpenTalk2HTML-NotMD
• npm: https://www.npmjs.com/package/@aimino/opentalk2html-notmd

---

License

GPL-3.0-only — See LICENSE in the repository.