How youssef-fd/offline-ai-agent works
This repository presents an offline developer assistant, but the checked-in runtime shown here is mainly a browser chat UI and Express proxy; the n8n workflow is referenced by the README but its implementation is not present in the tree. The evidenced path is browser input and text-file reading, through Express to the local n8n webhook, then back as a response. The README describes broader local-model, intent-routing, and code-editing stages, but their implementation and connections cannot be verified from the available files.
Main entry points: app.js Chat and uploads · server.js Express API · index.html Web page · schema.sql Postgres history
Components 9
Chat interface
- Chat and uploads
AI_Interface/public/app.js - Web page
AI_Interface/public/index.html
Web server
- Express API
AI_Interface/server.js
AI workflow
- n8n workflows
- Local model
- Code file workflows
- Intent routing
Session history
- Postgres history
AI_Interface/sql/schema.sql
Other components
- Developer
Connections 9
- Developer → Chat and uploads: uses
README.md - Chat and uploads → Express API: posts chat
AI_Interface/public/app.js - Chat and uploads → Express API: loads history
AI_Interface/public/app.js - Web page → Chat and uploads: loads script
- Express API → n8n workflows: forwards request
AI_Interface/server.js - n8n workflows → Local model: runs inference
README.md - n8n workflows → Intent routing: routes intent
README.md - n8n workflows → Code file workflows: handles files
README.md - Express API → Postgres history: queries history
AI_Interface/server.js