Developer-Native Marketing

Local Assistant Workflow

Use GTM Toolkit's built-in prompts with Cursor, Copilot, Claude Desktop, or any desktop assistant—no hosted API keys required.

Pick Your Assistant

Any local-first assistant that accepts text prompts works well.

• Cursor / VS Code + Copilot for in-editor research and fix suggestions.
• Claude Desktop when you want a dedicated workspace with memory.
• Custom internal assistants that accept structured prompts from JSON or markdown.
• Any chat-oriented tool that can browse URLs for competitor analysis runs.

Generate Prompts

Run the CLI and copy the prompt into your assistant of choice.

$ npx gtm-toolkit analyze --competitor https://example.com --output reports/competitor.json

🔁 Competitor benchmarking instructions ready

Instructions saved to reports/competitor.json

The saved file contains a LocalAIInstruction object with fields like title, objective, prompt, and suggestedSteps. Open it in your editor, copy the prompt, and paste it directly into your assistant.

{
  "title": "Competitor Benchmark Prompt",
  "objective": "Compare the competitor page against your positioning and highlight differentiation opportunities.",
  "prompt": "Visit https://example.com and evaluate the page...",
  "recommendedAssistants": ["GitHub Copilot", "Cursor", "Claude Desktop"],
  "suggestedSteps": [
    "Open the competitor URL in your browser.",
    "Feed the URL plus this prompt to your local AI assistant.",
    "Summarize the analysis in a short internal brief."
  ]
}

Customize gtm.config.js

Document team preferences so everyone follows the same workflow.

// gtm.config.js
module.exports = {
  ai: {
    assistant: 'local-ai',
    notes: 'Open prompts in Cursor or Copilot for follow-up research.',
  },
};

These fields are optional but helpful when sharing instructions with collaborators or CI pipelines. They never include API keys—just human-readable guidance.

Share & Review

Keep generated insights actionable for the rest of the team.

• Commit prompt JSON to a reports/ directory when you want a permanent record.
• Attach assistant output to pull requests so reviewers can see the recommendations that drove your changes.
• Tag prompts with the sprint or campaign they support to make later audits easier.

Security & Collaboration

Prompts are safe to store, but still treat competitive insights carefully.

• No secrets are required for local prompts, so you can run analyses without managing API keys.
• If competitor research contains sensitive data, save outputs in a private repository or shared drive.
• When collaborating across teams, document how prompts were used so marketing, product, and GTM stay aligned.