Alpi Marketer Admin

Manage companies, rooms, and agents

ENV:

🔄 Checking... - | -

🔄 DB: Checking... -

Create New Company

Company Context:

Context Token Limits Configuration:

Configure token limits for this company. Default values will be applied if not specified.

Token Limits Configuration

Total Context Tokens Limit: Total tokens allowed for complete agent context

LLM Provider: LLM provider (for future use)

Token Limit Percentage Split:

Company + Room + Agent Combined (%):

Agent Action History Context (%):

User Feedback Context (%):

History Summary Storage (%): Max stored history file size

Agent Token Limits

Post Summary Generation Limit (tokens): Max tokens for summarizing a post into one line (agent-specific)

Note: History condensing automatically uses the "History Summary Storage" limit from above as its generation limit, ensuring the condensed history always fits within the storage limit.

Configuration Preview

Select Company to Edit

Company Context:

Context Token Limits Configuration:

Token Limits Configuration

Complete Context Limit (tokens): Total tokens allowed for complete agent context

Token Limit Percentage Split:

Company + Room + Agent Combined (%):

Agent Action History Context (%):

User Feedback Context (%):

History Summary Storage (%): Max stored history file size

OpenAI API Limits

Post Summary Generation Limit (tokens): Max tokens OpenAI can generate when summarizing a post into one line

Note: History condensing automatically uses the "History Summary Storage" limit from above as its generation limit, ensuring the condensed history always fits within the storage limit.

Configuration Preview

Danger Zone

Create New ChatGPT-Room

Room Context:

Select Room to Edit

Room Context:

Danger Zone

Create New Agent

Company:

Room:

Agent Name:

Action Type:

Execution Type:

Send email notifications when posts are generated

Immediate Execution (Manual Trigger)

This agent will not run on a schedule. Posts will only be generated when manually triggered via the API endpoint: POST /load_social_media_posts_immediate?chat_id=YOUR_CHAT_ID

No scheduling configuration is required. The agent will use the timezone setting from email notifications if needed.

One Time Execution

Date:

Time:

Timezone:

Daily Execution

Time:

Timezone:

Weekly Execution

Select Days:

Monday Tuesday Wednesday Thursday Friday Saturday Sunday

Time:

Timezone:

Biweekly Execution (Every 2 weeks)

Start Date:

Day of Week:

Time:

Timezone:

Monthly Execution

Monthly Type:

Start Date:

Day of Month (1-31):

Time:

Weekday:

Time:

Week:

Weekday:

Time:

Timezone:

Business Days Only Execution

Start Date:

Time: Executes Monday through Friday only, excluding weekends

Timezone:

Custom Rate Schedule

Every:

Unit:

Note: Minimum interval is 5 minutes for EventBridge Scheduler.

Custom Schedule

Context Type:

Select Agent to Edit

Agent Information

Agent Name

Agent Description

Execution Type:

Send email notifications when posts are generated

Immediate Execution (Manual Trigger)

This agent will not run on a schedule. Posts will only be generated when manually triggered via the API endpoint: POST /load_social_media_posts_immediate?chat_id=YOUR_CHAT_ID

No scheduling configuration is required. The agent will use the timezone setting from email notifications if needed.

One Time Execution

Date:

Time:

Timezone:

Daily Execution

Time:

Timezone:

Weekly Execution

Select Days:

Monday Tuesday Wednesday Thursday Friday Saturday Sunday

Time:

Timezone:

Biweekly Execution (Every 2 weeks)

Start Date:

Day of Week:

Time:

Timezone:

Monthly Execution

Monthly Type:

Start Date:

Day of Month (1-31):

Time:

Day of Week:

Time:

Week of Month:

Day of Week:

Time:

Timezone:

End Date (optional):

Custom Rate Schedule

Every:

Unit:

Note: Minimum interval is 5 minutes for EventBridge Scheduler.

Custom Schedule

Monthly Execution (Nth weekday)

Day of Week:

Nth occurrence:

Time:

Timezone:

Business Days Only (Monday - Friday)

Time:

Timezone:

Agent History Access Configuration

Configure which agent histories this agent should consider when generating context. This affects the "agent action history" part of the baked context.

History Access Settings

Include this agent's own history Whether to include this agent's own action history in the context

Include history from other agents: Select other agents whose action histories should be included in this agent's context

Load config to see available agents...

Current Configuration Preview

Load config to see current settings...

ℹ️ How this works:

This configuration affects the "Agent Action History Context" part of the baked context
Histories from selected agents will be merged and summarized when generating context
This allows agents to learn from each other's past actions and results
Changes take effect immediately for new post generations

All Agent Contexts

Agent Context

Agent Action History Summary

This summary is automatically updated when agents execute actions. You can view or manually edit it here.

Single User Chat Feedback Extraction Context

Context for extracting feedback from individual user chat logs.

Single User Chat Feedback Integration Context

Context for integrating cumulative user feedback into the agent's context during post generation.

🔄 User Feedback Context (Baked)

The final baked user feedback context that combines integration context with cumulative user feedback. This is automatically generated and used during post generation.

🧪 Baked Context Debugging

Debug and preview the baked context generation process without modifying any data.

Danger Zone

🎯 Agent Monitor

📋 Executed Agents

Loading execution data...

File Browser

Agent Context & Configuration Files

Upload Custom Context

Chat Log Files

Generated Posts Repository

Select Agent for Post Generation

Context Information

ChatGPT logs will appear here...

🔄 Feedback Processing Pipeline

🧪 Post Generation Pipeline Testing

Test the intermediate steps before post generation to understand how context and feedback influence the final output.

Step 1: Context Processing Analysis

Click "Analyze Context Processing" to see how all contexts are merged and processed...

Step 2: User Feedback Summary

Click "Test Feedback Summary" to see how ChatGPT logs are analyzed for user feedback...

Step 3: Post Generation Logic Preview

Click "Preview Post Logic" to see how context + feedback will influence the final post...

Full Pipeline Result

Click "Run Full Pipeline" to execute the complete post generation process...

Generate Post

Send email notification

View Latest Posts

🧪 Batch Agent Testing

Batch 1

Execution Type:

Agent Type:

Number of Agents:

Time Until Execution:

Send email notifications

Test History & Status

No test runs yet

Create New Company

Token Limits Configuration

Token Limit Percentage Split:

Agent Token Limits

Configuration Preview

Select Company to Edit

Token Limits Configuration

Token Limit Percentage Split:

OpenAI API Limits

Configuration Preview

Danger Zone

Create New ChatGPT-Room

Select Room to Edit

Danger Zone

Create New Agent

Immediate Execution (Manual Trigger)

One Time Execution

Daily Execution

Weekly Execution

Biweekly Execution (Every 2 weeks)

Monthly Execution

Business Days Only Execution

Custom Rate Schedule

Custom Schedule

Select Agent to Edit

Agent Information

Agent Name

Agent Description

Immediate Execution (Manual Trigger)

One Time Execution

Daily Execution

Weekly Execution

Biweekly Execution (Every 2 weeks)

Monthly Execution

Custom Rate Schedule

Custom Schedule

Monthly Execution (Nth weekday)

Business Days Only (Monday - Friday)

Agent History Access Configuration

History Access Settings

Current Configuration Preview

All Agent Contexts

Agent Context

Agent Action History Summary

Single User Chat Feedback Extraction Context

Single User Chat Feedback Integration Context

🔄 User Feedback Context (Baked)

Raw Cumulative User Feedback:

🧪 Baked Context Debugging

Baking Steps Trace:

Full Baked Context Preview:

Danger Zone

🎯 Agent Monitor

📋 Executed Agents

File Browser

Agent Context & Configuration Files

Upload Custom Context

Chat Log Files

Generated Posts Repository

Select Agent for Post Generation

Context Information

🔄 Feedback Processing Pipeline

Pipeline Preview Results

Feedback Processing History

🧪 Post Generation Pipeline Testing

Step 1: Context Processing Analysis

Step 2: User Feedback Summary

Step 3: Post Generation Logic Preview

Full Pipeline Result

Generate Post

View Latest Posts

🧪 Batch Agent Testing

Batch 1

Current Test Run

Test History & Status