Trigger Inbox (Email Channel)
What is Trigger Inbox
Trigger Inbox (this document focuses on the default email channel first) is an external trigger capability. Users do not need to log in. They can send a natural-language email request (for example, "Send me last week's AI news"), and the system will either reuse an existing report or trigger template generation, then reply with the result by email.
Core Value:
- 🚀 Zero Barrier - No login required, just send an email
- 💬 Natural Language - Describe needs in everyday language, no commands to remember
- ⚡ Instant Response - Existing reports sent within 1 minute, generated reports in 5-15 minutes
- 📧 High-Frequency Channel - Leverage email as a high-frequency tool to increase engagement
Typical Scenarios:
- An investor sends an email in the morning: "Send me yesterday's AI news", system matches last night's auto-generated report and delivers within 1 minute
- A CEO sends an email on a plane: "Latest blockchain trends from last week", system finds matching template, triggers generation, and sends results in 15 minutes
- A product manager sends: "Show me competitor analysis", system replies "No available data" (no matching template)
PlaceholderEmail trigger workflow diagram - User sends email → System parses → Match/Generate → Reply email
Email Whitelist Configuration
For security and to prevent abuse, only emails from whitelisted addresses will be processed.
Default Configuration
Upon registration, the system automatically adds your registration email to the whitelist.
Example:
- Registration email: [email protected]
- Default whitelist: ["[email protected]"]
Add More Emails
You can add more email addresses (e.g., work email, personal email) in the Settings page, up to 5 addresses.
Configuration Steps:
- Log in and go to Settings page
- Find the Email Whitelist section
- Enter one email address per line
- Click Save
Configuration Example:
ℹ️ Note: Only emails from whitelisted addresses will be processed, others will be ignored.
Sending Request Emails
System Email Address
System email address: [email protected] (receives all user requests)
Email Format
Subject: Any (system only reads body content)
Body: Describe your needs in natural language
Example 1: Request existing report
Send me yesterday's AI news
Example 2: Trigger template generation
Latest blockchain trends from last week
Example 3: Specify topic and keywords
Show me recent AI investment and financing news
💡 Tip: More specific requests have higher match rates. Try to include: time range (e.g., "yesterday", "last week") + topic (e.g., "AI news", "blockchain").
PlaceholderEmail body example screenshots - showing 3 different types of request emails
Workflow
Step 1: Parse Intent (50 credits)
System uses LLM to parse your natural language request and extract:
- Time Range: "yesterday", "last week", "last month"
- Topic: "AI news", "blockchain trends"
- Keywords: ["AI", "news", "artificial intelligence"]
Example:
- Input: "Send me last week's AI news"
- Parsed result:
- Time range: 7 days
- Topic: AI news
- Keywords: ["AI", "news"]
Step 2: Match Existing Reports (20 credits)
System queries your reports from the last 7 days and uses LLM to determine if they match your needs.
Match success → Send report directly → Complete (Total cost: 70 credits)
Match failed → Proceed to Step 3
Step 3: Match Triggerable Templates (20 credits)
System queries all your report templates (that allow external triggers) and uses LLM to determine the best match.
Match success → Trigger report generation → Wait 5-15 minutes → Send report → Complete
Match failed → Reply "No available data" → Complete (Total cost: 90 credits)
Step 4: Send Results (10 credits)
- Success: Email contains report content (Markdown format)
- Failure: Email explains reason (e.g., "No available data", "Insufficient balance")
Progress Notifications
When triggering template generation, the system sends progress notifications at these intervals (free, no charges):
| Time | Notification Content |
|---|---|
| 5 min | "Your request is being processed, please wait... (waited 5 minutes)" |
| 10 min | "Your request is being processed, please wait... (waited 10 minutes)" |
| 15 min | "Your request is being processed, please wait... (waited 15 minutes)" |
| 20 min | "Your request is being processed, please wait... (waited 20 minutes)" |
| 25 min | "Your request is being processed, please wait... (waited 25 minutes)" |
| 30 min | "Your request is about to timeout..." |
Timeout Handling:
- If not completed within 30 minutes, system sends timeout notification (free)
- Status becomes TIMEOUT, no further processing
PlaceholderProgress notification email screenshot - showing notifications at different time intervals
Cost Breakdown
| Operation | Cost (Credits) | Description |
|---|---|---|
| Intent Parsing | 50 | Fixed per request |
| Match Existing Reports | 20 | Fixed per request |
| Match Templates | 20 | Only when no existing report |
| Generate Report | Per template config | Only when template matched |
| Send Email | 10 | Fixed per email |
| Progress Notifications | 0 | Free |
| Timeout Notification | 0 | Free |
Cost Examples:
Scenario 1: Existing Report Available
Intent parsing: 50 credits + Match report: 20 credits + Send email: 10 credits = Total cost: 80 credits
Scenario 2: Trigger Template Generation
Intent parsing: 50 credits + Match report: 20 credits (no match) + Match template: 20 credits + Generate report: 200 credits (assuming GPT-4 template) + Send email: 10 credits = Total cost: 300 credits
Scenario 3: No Match
Intent parsing: 50 credits + Match report: 20 credits (no match) + Match template: 20 credits (no match) + Send "No available data" notification: 0 credits (free) = Total cost: 90 credits
Usage Limits
Daily Quota
Maximum 10 requests per user per day (configurable in system settings).
Balance Requirement
Before initiating a request, system checks if balance is sufficient (at least 100 credits).
Duplicate Request Detection
Same user + same request content + within 1 hour → System replies "You have a pending request, please wait...", no new request created.
Best Practices
✅ Provide Specific Time Ranges
Recommended:
- "Send me yesterday's AI news"
- "Latest blockchain trends from last week"
- "Last month's investment report"
Not Recommended:
- "Send me AI news" (vague time range)
✅ Specify Clear Topics and Domains
Recommended:
- "AI investment and financing news"
- "Blockchain technology development trends"
- "Cloud computing industry news"
Not Recommended:
- "Send me a report" (unclear topic)
✅ Use Natural Language
System supports natural language, no need to memorize commands:
- "Send me recent AI news" ✅
- "Show me AI news" ✅
- "I want AI news" ✅
✅ Configure Template Tags
Add tags (e.g., "AI", "news", "weekly") to report templates to improve match rates.
Configuration Steps:
- Go to Reports → Select template → Edit
- Find Template Tags section
- Add relevant tags (e.g., "AI", "news", "weekly")
- Check Allow External Trigger
- Save
FAQ
1. Why didn't I get a reply to my request?
Possible Reasons:
- Sender email not in whitelist → Add it in settings
- Insufficient balance → Top up credits
- System email issue → Contact technical support
2. How to improve match success rate?
Suggestions:
- Provide specific time ranges and topics
- Add tags to report templates (e.g., "AI", "news")
- Ensure templates have "Allow External Trigger" checked
3. How long does report generation take?
Time Estimates:
- Existing report: Within 1 minute
- Trigger template generation: 5-15 minutes (depends on data sources and template complexity)
- Timeout: 30 minutes
4. How to view historical request records?
Log in → Task Logs → Filter task type inbound_request
Next Steps
- Report Generation - Learn about report template design
- Knowledge Base Management - Learn about data source configuration
- Credits & Logs - View cost details
- FAQ - More help information