Add vector DB abstraction supporting pgvector and ChromaDB, plus migration testing tool and documentation

Implement category-based pre-filtering, token-based history pruning, and output token cap adjustments

Enhance security, input sanitization, and request validation in app.py

Add README files for Mac and Linux

Update assistant persona to Mintoak Website Assistant, add evaluation scripts and update data

Add Mintoak overall culture query enhancement rule

Limit sources returned to top 1-2 and add scraping scripts

Implement systematic semantic relevance and domain-context filtering

Sort sources by priority (products > blogs > other)

Fix employee count hallucinations and expand fallback triggers

Fix technology head routing and add corporate M&A out-of-scope guardrails

Update leadership information, separate CTO query rules, and rebuild DB

Update KB addresses, dynamic aggregation, and increase chunk size

Update Knowledge Base with scraped non-blog/non-product pages and add scraper agent skill

Enhance retrieval filters, add career/address rules, increase max chunk size, and rebuild data

Update RAG instructions for B2B merchant engagement and add evaluation report

fix: allow address, contact, and office location keywords in out-of-scope check

Upload adapters/mintoak/adapters.safetensors with huggingface_hub

Upload adapters/mintoak/adapter_config.json with huggingface_hub

feat: integrate fine-tuned Qwen 2.5 1.5B persona and evaluator parity

Update app.py and chat_server.py

formated the output

improved chat

bug solved

Your commit message here

Update codebase and configure automatic environment switcher (excluding MLX adapters)

Update application files and add launch script

feat: implement config-driven query enhancement registry and sync all workspaces

Update app.py with query optimizer and add ask.py CLI helper

Sync frontend lead capture form and endpoints to Space

Deploy UI update like GPT and chat histroy

Optimize frontend HTML rendering by throttling markdown parse

Revert "Optimize PyTorch CPU thread count to prevent core thrashing"

Optimize PyTorch CPU thread count to prevent core thrashing

Align max_new_tokens to 256

Upgrade to Qwen 2.5 1.5B Instruct GPU model with PyTorch

Configure explicit thread count for CPU to prevent thrashing

Fix python f-string syntax error

Use pre-compiled CPU wheels for llama-cpp-python

Fix exit code 143 by removing GGUF pre-download from build stage

Optimize with GGUF CPU engine and fix catalog injection

Deploy RAG chatbot with auto-population

initial commit