Think of a folder of ready answers. The AI fetches relevant chunks before speaking, without a base, it only uses instructions and the current conversation.
What is it?
Each workspace has its own bases. You add documents (hours, prices, policies, FAQ); the system chunks, indexes, and the assistant queries them at question time (RAG).
In the dashboard
Automations → Bases → create base → add documents → link to an assistant.
Document
Title + text (sourceText). Status after save:
Chunking (optional)
Best practices
- Short topics with clear titles
- Update when price or policy changes
- Multiple bases by theme (sales, support)
- Test real questions on the assistant before going live
Not the operator Aura copilot. These bases are your workspace content for your customers via the assistant.
v1 API
Knowledge bases are managed in the dashboard (Automations → Bases). There are no public /v1/ routes for bases or documents in this documentation.
Next steps
