Pre-requisites
Before running this project, make sure you have the following components set up.
1Docker (Required)
This project runs entirely on Docker.
One-liner Docker installation:
docs.docker.com/get-dockerVerify installation:
docker --version
docker compose version2Google AI Credits (Required)
This project uses Google Gemini & Veo for AI generation.
Veo 3.1
Video / Reel generation
Imagen
Nano Banana models
Gemini 2.5
Prompt enhancement
Without credits, video generation will fail with rate-limit or quota errors. Ensure billing is enabled in Google Cloud.
3Kestra Setup (Orchestration)
Kestra is used to orchestrate AI generation, media uploads, and publishing.
Copy the docker-compose.yml from the GitHub repository:
How to Set Up (Step-by-Step)
Facebook & Instagram Setup
- Create a Facebook account
- Create a Facebook Page
- Create an Instagram Business Account (Must be PUBLIC)
- Connect the Instagram account to the Facebook Page
Meta Developer Setup
Go to developers.facebook.com and log in.
- Create a new app
- Select Other as app type
- Add Instagram product
- Click Instagram Setup
Generate Access Token
In the Graph API Explorer, select the following permissions:
Token Format check:
EAA...Base64 Encode Secrets
Encode your keys before adding them to docker-compose.
echo "VALUE_HERE" | base64Features
AI Content Automation
- Prompt to Image (IG Feed)
- Prompt to Reel (IG Reels)
- Automatic Prompt Enhancement
Event-Driven Automation
- Webhook-based comments
- Auto-reply workflows
- Moderation automation
Authorization & Security
Enterprise-grade security standards built-in.
Architecture
Sento leverages a containerized microservices architecture orchestrated by Kestra, ensuring scalability and isolation.
Docker Host
Kestra Engine
Orchestrator
Automation Flows
1. Content Generation Flow
How a user prompt becomes an Instagram post.
2. Comment Moderation
Real-time handling of user interactions.