askmdb turns natural language into MongoDB queries with tuned Qwen models
The askmdb creator released an open-source Linux CLI that fine-tunes Qwen models to translate natural-language questions into MongoDB aggregation pipelines with one-line explanations. The single binary combines llama.cpp with the official MongoDB C driver, requires confirmation before execution, and reportedly performs all processing locally. In the creator’s 500-question English and Spanish evaluation, the 9B version achieved 91.6% accuracy, while multi-collection $lookup joins remained the main failure area.
The project provides code, Apache-2.0 GGUF model weights, a deterministic pipeline verifier, and a detailed evaluation, making its locally runnable text-to-query approach independently testable.
#askmdb#Qwen#MongoDB#llama.cpp#MongoDB C Driver#Qwen3.5-9B#Qwen3.5-0.8B#Gemma 4 E4B-it#Gemma 4 E2B-it