Brand Roadie 🎸

Inspiration

I'm a singer-songwriter with three albums. About a month ago I spent weeks reading marketing books and using AI to finally understand my own brand — my values, my audience, my positioning.

It worked. But it was hard.

Brand Roadie is the tool I wish I had.

I believe every independent artist deserves a strategy — not just the ones with managers or marketing budgets. This started as something I built for myself. I think it can help thousands of artists worldwide.

What it does

Brand Roadie is a strategic co-pilot for independent musicians.

It guides artists through a conversational AI chatbot to define their brand DNA — values, aesthetic, target audience, and SMART goals. From that foundation, it generates two key outputs:

• 🗓️ Operational Calendar — a structured release timeline
• 🚀 Launch Playbook — a concrete action plan with content, promo milestones, and release strategy tailored to your brand

But the real differentiator is the personal MCP server. Every artist gets an endpoint they can connect to Claude, ChatGPT, or any MCP-compatible client — and talk directly to their own data in natural language.

The MCP server exposes 6 tools: get_artist_identity, get_tracks, get_releases, get_smart_goals, get_buyer_personas, and get_full_profile.

Authentication is handled via API keys (hashed with SHA-256), generated and revoked from the artist's dashboard. Compatible with Claude Desktop via mcp-remote:

{
  "mcpServers": {
    "brand-roadie": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://brand-roadie.vercel.app/api/mcp/mcp",
        "--header",
        "Authorization: Bearer ${BRAND_ROADIE_API_KEY}"
      ]
    }
  }
}

"What's my next release?" "Am I on track with my Spotify goal?" "What's my brand positioning?"

How we built it

• v0 — UI generation and rapid iteration
• Vercel AI SDK — chatbot and AI logic
• Supabase — auth and artist data persistence
• Custom MCP Server — built with mcp-handler, deployed on Vercel at /api/mcp/[transport], exposing each artist's data as private, authenticated tools
• Vercel AI Gateway — multi-model access

Challenges

Building per-artist MCP authentication was the hardest part — each artist needs a scoped API key (hashed with SHA-256) so their data stays completely private when connecting from external clients like Claude or ChatGPT.

Designing a chatbot flow that extracts structured brand DNA from open-ended conversation was also non-trivial. Getting the AI to ask the right questions, in the right order, and persist that into a usable schema took several iterations.

What we learned

MCP is not just for developers. When you expose the right tools, non-technical users — like musicians — can get real value from AI clients they already use. The protocol is a distribution channel.

And the best products come from problems you've lived yourself. I wasn't building for a hypothetical user. I was building for me, a month ago, sitting with a stack of marketing books trying to figure out who I was as an artist.

What's next

• Spotify integration to enrich brand analysis with real listening data
• Collaboration features for artists working with managers or small labels
• Help every bedroom artist compete like they have a full team behind them

Built With

• ai-sdk-6
• mcp
• next
• react
• shadcn/ui
• supabase
• supabase-(auth-+-postgres)
• tailwind-css