AI workflow MCP-native Claude · ChatGPT · Gemini · MCP
AI Proposal Drafter
Paste a messy discovery call and get back a first-draft Statement of Work in your studio's voice, with every vague part of the call flagged as a question before it drafts a single line.
What it does
- Runs on Claude, ChatGPT or Gemini.
- Reads call transcripts, email threads, or written notes.
- Builds a rate card with you (if you don't have one), using public benchmark sources.
The Breakdown
The build steps, prompt architecture, and a worked walkthrough. Open to members.
01 Gap Analysis runs before the draft, not after
02 The five locked sections it writes into
03 The scope filter: what's in the fee, what's an add-on
04 Where the price comes from, and where it refuses to guess
05 A worked run, and the voice it borrows
BECOME A MEMBER TO UNLOCK_
Become a memberHow to install
Every workflow runs on the AI you already use. Pick your platform; each setup takes a couple of minutes.
01 Claude Code MCP · recommended
One-line install. Registers as a project-scoped MCP server. No restart needed.
- Unzip the bundle into your project root.
- Run the install command — Claude Code adds the server to .mcp.json and reloads automatically.
- Type /mcp to confirm the connection. Say "draft an SOW from this transcript" to run.
claude mcp add ai-proposal-drafter --command node --args ./ai-proposal-drafter/mcp/server.js 02 Claude Desktop MCP-native
Paste a server entry into Claude Desktop config and the skill loads on next launch.
- Open claude_desktop_config.json (macOS: ~/Library/Application Support/Claude/, Windows: %APPDATA%\Claude\).
- Add the snippet below under mcpServers.
- Quit and reopen Claude Desktop. The skill appears in the tools tray.
{
"mcpServers": {
"ai-proposal-drafter": {
"command": "node",
"args": ["/absolute/path/to/ai-proposal-drafter/mcp/server.js"]
}
}
} 03 Claude.ai Projects Paste-in
Install as a Project: SKILL.md → Project Instructions, knowledge → Project Knowledge.
- Create or open a Project on claude.ai.
- Paste SKILL.md into Project Instructions.
- Upload every file from /knowledge/ into Project Knowledge.
- Open a chat in the Project. Say "draft an SOW from this transcript".
04 ChatGPT Project or Custom GPT
Paste-in via Project files, or build a Custom GPT for the team picker.
- In ChatGPT, go to Explore GPTs → Create.
- Paste SKILL.md into Instructions.
- Upload /knowledge/ files into the Knowledge field.
- Save. Open the GPT and say "draft an SOW from this transcript".
05 Gemini Gem
Create a Gem with the skill brain in its instructions field.
- Open Gemini → Gems → Create a new Gem.
- Paste SKILL.md into the instructions field.
- Upload the knowledge files.
- Save and open the Gem. Say "draft an SOW from this transcript".
Stacks with
These AI workflows are designed to chain together. Run one's output through another and you get a real production pipeline.
- AI Editor In Chief
Already use AI Editor In Chief from MWM Issue 02? Point AI Proposal Drafter at your voice-fingerprint.md and the SOW prose is shaped through your rule set automatically.
-
No fingerprint? AI Proposal Drafter uses your tone setting (Formal/Corporate, Collaborative/Creative, your own description, or a tone summary mirrored from a past SOW you paste). You can wire up the fingerprint any time later.