Nobody has to use the chat. Every interview can be taken in whatever form suits the person, and the result is the same: answers land in the Inventory and on the Process Map, and the person confirms what kawaru.app reads back.
| Way | Suits | How it works |
|---|---|---|
| Chat | Most people | kawaru.app asks, the person types, kawaru.app asks back. Short answers or one long message both work. |
| Voice input | People who would rather talk than type | The person dictates answers in the chat instead of typing them. |
| Voice interviewer | People who do not want to read or write at all | The interview runs as a spoken conversation, and can be held as a meeting. |
| Own AI assistant | Teams already working in Claude or ChatGPT | The interview runs inside that assistant through MCP, without opening kawaru.app. |
| Documents | Processes that are already written down | Procedures, checklists, policies, old diagrams and spreadsheets are attached, and kawaru.app reads them first. The interview then covers only the gaps. |
| Recordings | People who do not want to talk to a chat at all | A recording of a regular interview, a workshop or a walkthrough is attached. kawaru.app transcribes it and works from the transcript. |
What can be attached
- Text and documents:
.md,.markdown,.txt,.docx,.pdf, including scans. - Images:
.png,.jpg,.jpeg,.webp. - Diagrams:
.vsdxfrom Visio,.drawio, and.bpmnor.xmlfrom Signavio, Camunda, Bizagi or ARIS. kawaru.app redraws them as its own map, and the person confirms it. - Spreadsheets:
.xlsx,.csv. - Audio:
.mp3,.m4a,.wav,.webm,.ogg,.flac. Video is not supported yet.
Mixing them
The ways can be mixed. A process owner can upload an existing procedure, add a recording of last week's workshop, and fill the gaps by voice. A consultant can run interviews as usual and upload the recordings instead of asking people to use the chat.