# Aidbow

> Aidbow — AI, aimed. The creative operating system for AI agents: Demand → Brainstorm → Overview → Workflow.

## Documentation
- [DBOW explained: Demand → Brainstorm → Overview → Workflow](https://aidbow.com/docs/dbow-explained.html): A stage-by-stage tour of the DBOW loop — what each step consumes, what it produces, and why the contract between stages is what makes Aidbow composable.
- [The P-metrics: scoring creative ideas objectively](https://aidbow.com/docs/p-metrics.html): The P-metrics score creative ideas on seven dimensions so brainstorming stops being a matter of taste. Here is what each dimension measures and how to use the scores.
- [Quick Start: from a raw idea to a workflow in 10 minutes](https://aidbow.com/docs/quick-start.html): Install Aidbow, run the DBOW loop on a real demand, and get from a fuzzy idea to a shippable workflow in about ten minutes.
- [What is Aidbow? The creative operating system for AI agents](https://aidbow.com/docs/what-is-aidbow.html): Aidbow is the open-core creative operating system for AI agents — it gives them a right brain with the DBOW loop: Demand, Brainstorm, Overview, Workflow.

## Tips & Recipes
- [How to run a two-round AI brainstorm (and why the second round matters)](https://aidbow.com/tips/two-round-brainstorm.html): Round one is blind divergence; round two is cross-pollination. Running them separately is the cheapest way to get genuinely different ideas instead of variations on the first one.

## Q&A
- [Do AI agents actually improve creative output?](https://aidbow.com/qa/agents-improve-creativity.html): Agents are strong at divergent generation and weak at judgment. Used with the right process, they raise the floor of creative work — but they do not replace the human call on what to build.
- [Aidbow vs a plain ChatGPT prompt: what's the difference?](https://aidbow.com/qa/aidbow-vs-chatgpt.html): A single chatbot prompt collapses four different mental modes into one pass. Aidbow separates them into the DBOW loop so the output is aimed, scored and auditable.
- [Is the open-source version enough, or do I need the precision layer?](https://aidbow.com/qa/open-source-vs-precision.html): 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.

## Resources
- [Toolbox: prompts, templates and checklists](https://aidbow.com/resources/toolbox.html): A starter kit of the artifacts that make creative AI work repeatable — a demand brief template, an idea card, a scoring rubric and an overview checklist.
- [Lessons: what breaks when agents brainstorm](https://aidbow.com/resources/what-breaks-when-agents-brainstorm.html): Five failure modes we hit running multi-agent brainstorming in practice — fake consensus, hallucinated sources, template repetition, scoring drift and the originality tax — and the concrete fix for each.

## Engineering Notes
- [Anti-hallucination: mandatory provenance and explicit conflicts](https://aidbow.com/notes/anti-hallucination-provenance.html): Hallucination in a summarizer is not a model bug you prompt away — it is a missing constraint. Two rules — mandatory provenance and explicit conflict flags — cut it down more than any prompt tweak.
- [Multi-agent collaboration: blind rounds and convergence](https://aidbow.com/notes/multi-agent-blind-rounds.html): How to structure multi-agent brainstorming so agents produce independent ideas before they see each other's work — and what to do when they converge anyway.
- [Creative evaluation: splitting novelty into R1 and R2](https://aidbow.com/notes/novelty-split-r1-r2.html): Treating originality as one score breaks the moment a second round exists. Splitting it into R1 (independent novelty) and R2 (incremental value over the existing set) fixes the metric.