Overview
Sessions β last 14 days
Knowledge base
Knowledge Base
Crawled from the website, chunked, embedded with Gemini and stored in the vector store.
RAG source websites
Add each website you want the assistant to use as knowledge. The crawler stays within each website, respects robots.txt, and indexes saved sources together.
Search vectors
Indexed pages
| URL | Title | Chunks | Status | Last crawled |
|---|
Crawl history
| Started | Status | Found | Indexed | Skipped | Failed | Chunks |
|---|
Conversations
| Started | Origin | Status | Duration | Msgs | Latency |
|---|
Leads
Captured by the save_lead_info tool (also mirrored to the lead CSV).
| Updated | Name | Interest | Budget | Mobile | Details |
|---|
Logs
| Time | Source | Level | Message |
|---|
Settings
Assistant
Environment (read-only β set in Coolify and redeploy)
Security
These controls are enforced immediately for new widget requests and calls. Credentials remain server-only environment variables.
Website access
Only these hosts can open the widget. Use one host or origin per line, such as textileprojects.in, https://www.textileprojects.in, or *.textileprojects.in. Use * only while developing.
Call limits and timeouts
Use -1 to turn off a limit or silence cutoff. Positive values are measured in seconds, except the two daily call limits.
Conversation protection
Embed Code
Paste this keyless snippet into the <head> of your approved website. The widget API key stays on the server and is never placed in page source.
Loadingβ¦
Origin-bound widget keys
Create one key for each website. Its explicit host list resolves the secure, keyless embed; do not use * here.
| Key | Name | Origins | Created |
|---|