Is the open-source version enough, or do I need the precision layer?
The open-source framework is complete enough to run the whole DBOW loop. The precision layer adds project-tuned scoring, refined prompts and a production harness — it is an upgrade, not a prerequisite.
TL;DR — The open-source framework (MIT) runs the entire DBOW loop with no strings: install it, run the demo, use it in production. The precision layer is for teams who want project-tuned scoring and refined prompts without maintaining them.
What you get open source
Everything needed to run the loop:
- The DBOW orchestration (demand → brainstorm → overview → workflow)
- A baseline P-metrics engine and prompt pack
- Context providers and a review UI
- An offline demo that runs with no API key
That is not a teaser. You can build on it.
What the precision layer adds
It implements the same interfaces and extends the open base without forking it:
| Area | Open base | Precision layer |
|---|---|---|
| P-metrics | baseline dimensions | project-tuned rubric, split novelty (R1/R2) |
| Prompts | general templates | refined, anti-hallucination-grade |
| Harness | swappable steps | production step contracts |
| Depth scoring | self-reported | computed from actual substance |
Because it plugs in through the documented extension points, you keep tracking upstream releases instead of maintaining a fork.
Which do you need?
- Exploring, learning, or building your own: start open source.
- Running creative work at volume and wanting the sharp edges pre-tuned: the precision layer saves you the tuning.
The interfaces are public — see the open-source spec.