You can use google-generative AI api , for it preferably
Add a lightweight interpreter that converts natural text instructions into corresponding bot commands.
This does not require full AI — a rule-based intention engine is acceptable.
Example inputs:
"bot follow me" → .follow
"drop everything" → .throwall
Context & Approach Guidance
Focus on:
Identifying intent keywords (follow, mine, throw, stop)
Mapping them to command functions
Handling ambiguous instructions politely
Adding optional AI model support via .env key
Architecture considerations:
Should interpreter run before or after manual command parser?
Should feedback always show the interpreted command first?
Acceptance Criteria
Normal commands still work unchanged
Without AI key: fallback to rule-based routing
Outputs interpreted command before execution
Handles unknown phrases with friendly fallback
Labels: hard, AI, architecture, feature
You can use google-generative AI api , for it preferably
Add a lightweight interpreter that converts natural text instructions into corresponding bot commands.
This does not require full AI — a rule-based intention engine is acceptable.
Example inputs:
"bot follow me" → .follow
"drop everything" → .throwall
Context & Approach Guidance
Focus on:
Identifying intent keywords (follow, mine, throw, stop)
Mapping them to command functions
Handling ambiguous instructions politely
Adding optional AI model support via .env key
Architecture considerations:
Should interpreter run before or after manual command parser?
Should feedback always show the interpreted command first?
Acceptance Criteria
Normal commands still work unchanged
Without AI key: fallback to rule-based routing
Outputs interpreted command before execution
Handles unknown phrases with friendly fallback
Labels: hard, AI, architecture, feature