Switch from minimax/minimax-m3:free to inclusionai/ling-3.0-flash as the
default AI model for insight generation, aligning with the updated AI
provider configuration.
Implement local database storage to automatically save all sent insights
with metadata including topic, content, AI provider, model, and timestamp.
Add database module with initialization and query functions, integrate
with main job flow, and update documentation with database schema details.
Remove StepFun provider support and consolidate AI configuration under
DEFAULT_AI_PROVIDER variable supporting openrouter, llamacpp, and openai.
Update .env, README.md, and main.py to use the new provider-based structure.
Add "Qualidade Total na Empresa" to topics list.
Implement message splitting for messages exceeding Telegram's 4096 character
limit. Update header branding, increase max_tokens to 4000, adjust END_HOUR
to 20, update default AI model, and revise documentation to reflect channel
setup instead of groups.
Adds comprehensive documentation for creating a Telegram bot, configuring
a group, and setting permissions. Updates TELEGRAM_CHAT_ID in .env to
the correct group identifier for automated messaging.
Implements ThorAgentIA bot that generates and sends scheduled insights
via Telegram using configurable AI providers. Includes environment
configuration, scheduling, and setup automation.
marco