Repository navigation
features
Active contributors: ferdiiskandar
MyPrompt's product behaviour is carried by four feature areas under the repository root.
- Transform is a deterministic, no-network prompt scaffold. It detects intent, picks a persona, and compiles the raw prompt into the shape a specific target model expects (Claude XML, Codex acceptance criteria, Gemini schema, Grok evidence framing).
- Optimizer is LLM-backed and produces one of two output kinds: a six-section Super Prompt or an eight-section Coding Brief checked by validator V1–V14.
- Evaluator scores a prompt with an LLM-as-a-judge across four weighted dimensions.
- Workspace and library covers local drafts, recent runs, benchmarks, the Prisma-backed prompt library, and the template catalog.
All four run in the desktop main process.
Open-source prompt engineering and multi-LLM tooling by Sentra Artificial Intelligence.
MyPrompt develops practical approaches to prompt engineering, multi-LLM optimisation, reusable prompt systems, and AI-native workflows — with an emphasis on structured, interoperable, and real-world AI use.
Built in Indonesia as part of the Sentra Artificial Intelligence ecosystem.
Sentra Artificial Intelligence · Source Repository · Official Website
Dr Ferdi Iskandar — Creator & Maintainer
LinkedIn ·
ORCID ·
Hugging Face ·
Kaggle ·
Medium ·
Substack ·
X ·
Threads
MyPrompt · Sentra Artificial Intelligence · Indonesia