Skip to main content

Tech Stack & Web Modules

This document outlines the core web endpoints and the underlying infrastructure that powers the Sentra ecosystem.

Web Modules / Endpoints

The Sentra platform is distributed across several domain-specific microservices:

ServiceEndpointDescription
Sentra Landinglanding.bitgrit.aiMarketing and platform overview.
Sentra OIDC Authauth.woollet.org/registerUser registration and authentication gateway.
Sentra ID / Adminauth.bitgrit.netIssuer registry, Schema maintenance, VC issuance, and Blockchain integrations.
Sentra Walletwallet.bitgrit.netEnd-user interface for VC and Asset management.
Sentra v3 (Core App)sentra.bitgrit.aiThe primary Multi-tenant AI platform and management console.
Bitgrit AI Corebitgrit.comThe core AI competition and custom development platform.

Platform Infrastructure

1. Multi-Tenant Application Architecture

Sentra operates on a 1-to-Many architecture:

  • One Unified API Service: A centralized Python FastAPI backend processing all orchestration, LLM interfacing, and database queries.
  • Many Frontends: Multiple Next.js applications, each isolated by a unique client_id (Tenant). The unified backend securely multiplexes requests based on this identifier.

2. Document Processing & Transcription Service

An independent, asynchronous service dedicated to ingesting complex media:

  • Audio/Video Transcription: Automatically extracts text from uploaded audio and video files.
  • Data Extraction: Automatically parses large transcripts to populate predefined application forms and feeds the extracted text directly into the RAG vectorization pipeline.

3. Blockchain Layer (Avalanche)

The migration away from legacy systems utilizes Avalanche for its high-speed consensus and EVM compatibility.

  • DID Registry: Smart contracts handling did:ethr document anchoring.
  • Revocation Registry: High-throughput on-chain state management for credential validation.

4. Identity & Access (OIDC)

  • Built upon robust OpenID Connect standards.
  • Supports SIOPv2 (Self-Issued OpenID Provider v2) bridging Web2 authentication flows with Web3 DID paradigms.
  • Full integration with modern biometric standards (Passkeys, WebAuthn).

5. Data & Storage

  • Graph & Vector Databases: Utilizes Neo4j for complex knowledge graph relationships and Qdrant for high-speed Vector embeddings retrieval within the LLM RAG pipelines.
  • IPFS / Pluggable Cloud: Google Cloud Storage (GCS) and IPFS for raw document storage (PDFs, Audio, Video).
  • TEE (Trusted Execution Environments): Secure enclaves (e.g., Intel SGX / AWS Nitro) for processing sensitive formula-based disclosures.