R1 Code Assist + Cline / Windsurf - This AI Assistant can GET CUSTOM Context for YOUR AI Coder!

---

---

AI Summary

Video Summary

Creating an Advanced AI Coding Helper with Vector Shift and Deep Seek R1

• Objective: To create an AI coding assistant that can reference a custom knowledge base or perform web searches and generate a markdown file for use in coding or other contexts.
• Tools Used: Vector Shift platform with Deep Seek R1 model (reasoning pro model) and perplexity for search functionalities.

Steps to Create the AI Assistant

1. Sign Up and Access Pipeline:
   • Register on Vector Shift and navigate to the pipeline section.
2. Pipeline Configuration:
   • Start with an input node.
   • Connect the input to an LLM (Language Learning Model) using the perplexity reasoning pro model.
   • Add an output node for the results.
3. Incorporating Custom Knowledge Bases:
   • Drag in a knowledge base node.
   • Create or select an existing knowledge base.
   • Add PDFs, documents, or webpage URLs to the knowledge base.
   • Connect the knowledge base in the pipeline with variables for prompt and context.
4. Export and Integration:
   • Export the setup as an automation, chatbot, or API.
   • For terminal use, integrate the API into a Python program that saves output to a markdown file.
5. Running the Assistant:
   • Use the Python program to input prompts and receive markdown files with the required information.
6. Additional Configurations:
   • Reference custom documentation, GitHub issues, Google Drive data, etc.
   • Edit the markdown file as needed.
   • Configure workflows to suit specific needs, not limited to coding.

Conclusion

• The AI assistant can be used for various purposes, including coding, documentation, and complex workflows.
• Vector Shift’s drag-and-drop interface and integrations make it user-friendly.

Additional Notes

• The AI assistant can be integrated into applications like WhatsApp, Slack, or used directly via API.
• Users can edit the context markdown file to ensure accurate referencing in tools like Klein or Cursor.
• The AI can generate examples from non-existent documentation, showcasing the capabilities of the R1 model.

Feedback and Support

• Viewers are encouraged to share their thoughts and subscribe to the channel for more content.
• Support options include donations or joining the channel for perks.