Skip to content

Core Technologies

This page describes the technology stack and key dependencies that power the GamiBot platform.


Technology Stack

ComponentTechnologyPurpose
LMSMoodle (custom PHP plugins)Course and resource management
OrchestrationLangFlowWorkflow design and agent orchestration
Vector DatabaseQdrantSemantic search and vector storage
EmbeddingsConfigurable (OpenAI, local, etc.)Convert text to dense vectors
LLMGPT-4, Claude 3, or local alternativesAgent reasoning and generation
ORMPrismaDatabase abstraction (if using relational DB)
BackendNode.js + Express or PythonAPI layer and webhook handlers
File ProcessingUnstructured, PyPDF2, python-pptxExtract text from diverse formats

Key Dependencies

Moodle Plugins

  • local_gamibot_manager - Webhook dispatcher for file upload events

LangFlow Integration

  • Bundled Qdrant integration for vector operations
  • LLM connectors for multiple providers (OpenAI, Anthropic, local models)

Vector Operations

  • Qdrant client library - Python or Node.js SDK for vector operations
  • Collections support sharding and replication for scalability

Text Extraction Libraries

FormatLibraryNotes
PDFPyPDF2, pdfplumberOCR fallback available
PowerPointpython-pptxExtracts slides + speaker notes
ePubebooklibChapter parsing support
Wordpython-docxFull document support
Plain TextNativeUTF-8 encoding

Embedding Models

The platform supports configurable embedding models:

OpenAI Embeddings

json
{
  "model": "text-embedding-3-small",
  "dimensions": 1536,
  "max_tokens": 8191
}

Local Alternatives

  • Sentence Transformers - Self-hosted option for privacy
  • all-MiniLM-L6-v2 - Lightweight, efficient embeddings
  • BGE - Multilingual support

LLM Provider Options

Cloud Providers

ProviderModelBest For
OpenAIGPT-4, GPT-4oGeneral purpose, high quality
AnthropicClaude 3 Opus/SonnetLong context, safety
GoogleGeminiMultimodal, cost-effective

Local/Self-Hosted

ModelFrameworkNotes
Llama 3Ollama, vLLMOpen source, privacy
MistralOllama, vLLMEfficient, multilingual
Phi-3OllamaSmall, fast

Infrastructure Requirements

See the Deployment Architecture page for detailed infrastructure specifications.

Released under the MIT License.