Build Pangram
NOT REALLYreplaces $20/mosaves $240/yrback to the verdict
Before step 1
you will need
- Local Python with torch and transformers
- A couple of GB of disk for small model weights, CPU works but is slow
- Your own labelled samples of human and AI text if you want any idea of accuracy
Delivery order
Scaffold the smallest runnable application and document its commands.
done whenThe project starts from a documented command in a clean checkout.
Implement the primary data model and core workflow.
done whenThe main object can be created, read and changed end to end.
Add validation, safe failure states, and persistence.
done whenBad input is refused with a readable message and nothing is left corrupted.
Cover the critical path with automated tests.
done whenThe highest-risk behavior fails the suite when it breaks.
Exercise a clean install from the README and fix every missing step.
done whenA fresh clone reaches the first successful workflow using only the README.
That is the whole plan for Pangram. What it deliberately does not cover is below · check the gaps before you call it a replacement.
- Calibration: a real false positive rate you can quote, instead of a vibe
- Coverage of new models, which changes every few weeks whether you update or not
- Sentence-level and mixed-authorship detection rather than one blunt document score
- Any external credibility, since a self-built score persuades exactly zero teachers, editors or clients
- Throughput, batch uploads, API access and document parsing
Need the files? The project pack on the verdict page hands your agent the whole brief · more ai writing.