feat(pipe): add crm-pipe integration (issue #1073)

[Youngthug38]

description

/claim #1073

This PR implements the CRM Pipe requested in issue #1073. It creates an automated "Cheatsheet" for every person the user interacts with by analyzing audio transcripts.

Key Features:

• Cron Scheduling: Polls for new audio data every hour (configurable).
• Speaker Isolation: Groups transcripts by speaker_id to handle multiple conversations.
• LLM Extraction: Uses OpenAI (or local Ollama) to extract:
  • Psychometric Profile: Big Five personality traits & communication style.
  • Pain Points: Explicit frustrations/complaints.
  • Desires: Stated goals and wants.
• Local CRM: Persists profiles as human-readable Markdown files in ~/.screenpipe/pipes/crm-pipe/data/.
• Mock Testing: Includes a simulation server for easy verification.

related issue: #1073

how to test

I have included a Mock Server (src/mock-server.ts) so you can verify the logic immediately without needing to generate real audio data in the desktop app.

1. Navigate to the pipe directory: cd pipes/crm-pipe and run npm install.
2. Terminal A (Simulator): Run npx ts-node src/mock-server.ts (Starts a fake Screenpipe API on port 3030).
3. Terminal B (The Pipe): Run npm run build && npm start.
4. Verify: Check the data/ folder. A file named speaker_1.md will appear containing the extracted personality profile.

Demo Video:
https://youtu.be/xS_pws8iMnE

Screenshots

[github-actions]

🧪 testing bounty created!

a testing bounty has been created for this PR: view testing issue

testers will be awarded $20 each for providing quality test reports. please check the issue for testing requirements.

[louis030195]

hey @Youngthug38! screenpipe 2.2.0 is here! 🎉 check out the demo: https://x.com/louis030195/status/2014403576475615312

closing due to inactivity - CRM pipe is cool! if you'd like to continue, please open a fresh PR against main. would love your feedback on the new version!