This repository contains the Terraform configuration for the Memovee Tama application, which provides a comprehensive AI-powered conversational platform integrating with various services including Tama, Mistral, X.ai, and Elasticsearch. The system supports advanced features like personalized conversations, media handling, and intelligent message routing.
The Memovee Tama system is designed to create an intelligent conversational AI platform that can handle:
- Personalized conversations with user profiles
- Media-related interactions
- Message routing and classification
- Contextual awareness and memory
- Integration with external AI services
- Core memovee messaging module serving as the primary interface
- Reply generation chain with associated processors
- Configuration for handling reply generation with specific models and contexts
All models referenced via *_model_id, *_model_parameters, and tama_thought_processor resources:
| Function | Model | Temperature | Reasoning effort | Service tier |
|---|---|---|---|---|
| Memovee - Reply Generation | gpt-5-mini | 1.0 | low | |
| Router - Message Routing | gpt-5-mini | 1.0 | minimal | |
| Elasticsearch - Index Mapping Generation | gpt-5 | 1.0 | low | |
| Index Definition Generation | gpt-5 | 1.0 | low | |
| Movie DB - Generate Description | gpt-5-nano | 1.0 | low | flex |
| Movie DB - Generate Setting | gpt-5-nano | 1.0 | low | flex |
| Media Browsing - Tool Call | gpt-5-mini | 1.0 | minimal | |
| Media Detail - Tool Call | gpt-5 | 1.0 | minimal | |
| Person Browsing - Tool Call | gpt-5-mini | 1.0 | minimal | |
| Person Detail - Tool Call | gpt-5 | 1.0 | minimal | |
| Memovee - Artifact Creation | gpt-5-mini | 1.0 | low | |
| Basic - Check Profile Tooling | gpt-5 | 1.0 | minimal | |
| Basic - Upsert Profile Tooling | gpt-5 | 1.0 | minimal |
- Elasticsearch module for indexing and searching capabilities
- Personalization and prompt assembly spaces
The system is built with a modular architecture that allows for flexible extension and maintenance of different conversational functionalities.
- Main: Core memovee functionality and interfaces
- Basic: Basic conversation handling with off-topic, greeting, and introductory flows
- Media: Media-related conversation components
- Personalization: User-specific content management
- Prompt Assembly: Contextual information assembly
- Router: Message classification and routing
- Movie Database: Movie information integration
The following diagram shows the correct relationships between Tama spaces in the system, based on the actual tama_space_bridge resource definitions:
graph LR
A[Main Memovee Space]
B[Basic Conversation Space]
C[Media Conversation Space]
D[Personalization Space]
E[Prompt Assembly Space]
F[Movie DB Space]
G[Elasticsearch Space]
H[Memovee UI Space]
A -- "memovee-basic" --> B
A -- "memovee-media" --> C
A -- "memovee-ui" --> H
B -- "basic-conversation-personalization" --> D
B -- "basic-conversation-prompt-assembly" --> E
B -- "basic-conversation-memovee" --> A
C -- "media-conversation-to-movie-db" --> F
C -- "media-conversation-to-prompt-assembly" --> E
E -- "prompt-assembly-memovee" --> A
F -- "movie-db-elasticsearch" --> G
style A fill:#bbdefb stroke:#333
style B fill:#c8e6c9 stroke:#333
style C fill:#ffe0b2 stroke:#333
style D fill:#f8b6c0 stroke:#333
style E fill:#e1bee7 stroke:#333
style F fill:#d7ccc8 stroke:#333
style G fill:#fff9c4 stroke:#333
style H fill:#cfd8dc stroke:#333
For detailed information about the various components of the Memovee Tama system, please refer to the following documentation files:
- Main Module Documentation
- Movie Database Documentation
- Personalization Documentation
- Router Module Documentation
To get started with the Memovee Tama project, you'll need to:
- Set up your environment with the required Terraform version (v1.0.0+)
- Configure API keys for external services (Tama, Mistral, X.ai, Elasticsearch)
- Create a
.auto.tfvarsfile with your configuration values - Run
terraform init,terraform plan, andterraform apply