Spaces:
Running
Running
|
Download CONTEXT.md from IDKHowToCodeFr/vera-engine: direct link, hf CLI and curl.
- Browser
- Download file 1.14 kB
-
https://huggingface.co/spaces/IDKHowToCodeFr/vera-engine/resolve/main/CONTEXT.md
- Command line
-
hf download hf://spaces/IDKHowToCodeFr/vera-engine/CONTEXT.md
-
curl -L -o CONTEXT.md https://huggingface.co/spaces/IDKHowToCodeFr/vera-engine/resolve/main/CONTEXT.md
1.14 kB
| # Domain Context | |
| ## Glossary | |
| * **Vera**: AI bot β WhatsApp engagement β grow merchant biz. | |
| * **4-Context**: | |
| 1. `CategoryContext`: Vertical info (dentists/salons). | |
| 2. `MerchantContext`: Biz state (perf/identity). | |
| 3. `TriggerContext`: Event β convo start. | |
| 4. `CustomerContext`: Opt data (acting for merchant). | |
| * **Grounding**: 100% real metrics from req. 0 metric hallucination. | |
| * **Curiosity Hook**: Hook merchant when no perf data. | |
| * **Hinglish**: Hindi+En mix (60/40) β relate to IN merchants. | |
| ## Rules | |
| * **Len**: No hard cap. Keep terse/readable. | |
| * **CTA**: 1 clear action. YES/STOP or open. | |
| * **Tone**: Match category `voice` (peer/clinical). No promo. | |
| * **Lang**: Match merchant pref. | |
| * **Speed**: Res < 30s. | |
| ## Tech Strategy (50/50 Score) | |
| * **Distill**: `Context Distiller` β shrink 500KB JSON β 4 bullets. | |
| * **Router**: API req β `compose_with_fallback`. 4-tier chain (NVIDIA β Groq β Gemini β Local `llama-cpp-python` `qwen2.5-3b-instruct`). | |
| * **Format**: Schema force β pure JSON. | |
| * **Idempotent**: `/v1/healthz` β clear `contexts`. | |
| * **Hooks**: Loss aversion, peer proof, scarcity, binary commit. | |