Filter sessions by website/date-time, load and continue chat here.
Filters
Website
website_id
Timezone
From
To
Quick
—
Sessions
Paste session_id
—
Loaded Session Conversation
PDP - Product Insights
Focus: product-context Q&A testing
API Base
Environment
Website
website_id
user_id
session_id:—
product_id
Product Questions
AI Product Insights Chat
⚙️ Configuration
API Base
Env
Email
Password
Not logged in
Login calls POST /api/login and stores token locally. Configuration endpoints require Bearer token. Use the website picker below to load and edit a website’s configuration.
Website
website_id
Status—
Index
Search
0 changes
Save + Reprocess calls POST /api/reprocess-configuration and starts CONFIG_REINGESTION job.
—
Edit prompt text and JSON settings. Save writes to PUT /api/prompts.
prompt_id—
prompt
model_settings (JSON)
tools (JSON)
output_structure (JSON)
—
This edits agent configs (used by classifier + routed agents). Save writes to PUT /api/agents.
agent_id—
name
instructions
model_config (JSON)
tools (JSON)
output_structure (JSON)
custom_fields (JSON)
—
—
Manage website settings by category. Load existing category, edit JSON value, or add a new category.
Existing categories
Category
value (JSON)
—
Approves ingestion for the loaded website. This triggers a backend job via POST /api/approve-configuration.
website_id
—
website
—
—
Index Keys Editor
Full-window edit mode
Prompts Editor
Full-window edit mode
Agents Editor
Full-window edit mode
Add website setting category
Create a category with JSON value
Category
value (JSON)
—
Validate regex
—
regex_to_clean (JSON array)
—
—
New AI rules added to editor
Tip: “Preview next 10” pages through new samples.
Edit chunking config
—
chunking_config (JSON)
—
Saved via PUT /api/index-keys for this key.
Approve ingestion
—
This will start an ingestion job for the selected website. Please confirm you want to proceed.
website_id
—
website
—
I acknowledge this will trigger ingestion for this website.
—
Next steps: after approval, a job is created and queued. Use the returned job_id / task_id to track it in your backend logs/queue.