Gallery
Sentience
★ High quality
977
Sentience

Sentience

Hours 120h 41m

by oneonealex

Preface/Disclaimer: I'm submitting/shipping the initial version with a few features not implemented yet (there are some unused database tables in the unidb database), because I want to secure my spot at The Game™ in NYC. Once the project is completely done, I'll reship. This is a bot that monitors all messages on a discord server and processes them (extracting users' skills, emotions, and general knowledge about them (called 'facts' in the db)). The code Architecture looks like this: 1. Ingress bot: Pulls all messages from a discord server: Code lives at https://github.com/Creative-Difficulty/sentience-workers/tree/main/ingest-vestibule-retriever 2. Backend: Multiple microservices (currently "fact-extractor" and "topic-sorter", all microservices are also in the sentience-workers repo) each analyze the messages using LLMs and extract facts, activites and emotions, and sort them into topics. They then insert that data into the same database where the ingested discord messages are stored: unidb. 4. Sentience-Dashboard: A website that pulls from the database for cool visualizations and a dashboard of user information and LLM-inferred connections and facts, activities, detected emotions, and messages sorted into topics. There is also an S3 bucket used for storing message attachments, custom emojis and stickers. The unidb repo lives at https://github.com/Vimothy-s-Vestibule/unidb-schema and holds the postgres database schema and rust models that are used to store everything the bot collects and everything the processing microservices extract and infer from that. The project is dockerized, so it can be ran on one's own machine without much setup Note that there are some references to the word vestibule in numerous places in the project, this is because the discord server for which it was originally developed is called "The Vestibule" (there is no demo URL, since its a discord bot with a fancy AI analysis backend but there are setup instructions in the readme :))

Extra info
Submitted at 2026-05-10T20:59:46.070Z Approved at 2026-05-10T21:04:32.006Z
User submitted 126h 45m Reviewer approved 120h 41m Hackatime projects: [23519, 9575, 15035, 17261, 9464, 9465, 9466, 9467, 9572, 9492]
Raw project json
{
    "aasm_state": "approved",
    "ai_declaration": "I used AI sometimes to keep the rust structs up to date with the SQL files i wrote in unidb.\r\nAlso to understand all the events the discord API can emit, I used a bit of ChatGPT.\r\nTo help me with UI design on the dashboard (due to time constraints) and frontend work, as well as writing dockerfiles (I am not very good at that unfortunately) and trubleshooting I used the help of AI too.",
    "approved_at": "2026-05-10T21:04:32.006Z",
    "approved_seconds": 434479,
    "created_at": "2026-03-17T09:29:46.514Z",
    "demo_link": "https://github.com/Creative-Difficulty/sentience-workers/releases/tag/v1",
    "desc": "Preface/Disclaimer: I\u0027m submitting/shipping the initial version with a few features not implemented yet (there are some unused database tables in the unidb database), because I want to secure my spot at The Game\u2122 in NYC. Once the project is completely done, I\u0027ll reship.\r\n\r\nThis is a bot that monitors all messages on a discord server and processes them (extracting users\u0027 skills, emotions, and general knowledge about them (called \u0027facts\u0027 in the db)). \r\n\r\nThe code Architecture looks like this:\r\n1. Ingress bot: Pulls all messages from a discord server: Code lives at https://github.com/Creative-Difficulty/sentience-workers/tree/main/ingest-vestibule-retriever\r\n2. Backend: Multiple microservices (currently \"fact-extractor\" and \"topic-sorter\", all microservices are also in the sentience-workers repo) each analyze the messages using LLMs and extract facts, activites and emotions, and sort them into topics. They then insert that data into the same database where the ingested discord messages are stored: unidb.\r\n4. Sentience-Dashboard: A website that  pulls from the database for cool visualizations and a dashboard of user information and LLM-inferred connections and facts, activities, detected emotions, and messages sorted into topics.\r\n\r\nThere is also an S3 bucket used for storing message attachments, custom emojis and stickers.\r\n\r\nThe unidb repo lives at https://github.com/Vimothy-s-Vestibule/unidb-schema and holds the postgres database schema and rust models that are used to store everything the bot collects and everything the processing microservices extract and infer from that.\r\n\r\nThe project is dockerized, so it can be ran on one\u0027s own machine without much setup\r\n\r\nNote that there are some references to the word vestibule in numerous places in the project, this is because the discord server for which it was originally developed is called \"The Vestibule\"\r\n\r\n(there is no demo URL, since its a discord bot with a fancy AI analysis backend but there are setup instructions in the readme :))",
    "hackatime_projects": [
        23519,
        9575,
        15035,
        17261,
        9464,
        9465,
        9466,
        9467,
        9572,
        9492
    ],
    "high_quality": true,
    "id": 977,
    "real_approved_seconds": 434520,
    "rejected_at": "2026-05-10T20:46:09.982Z",
    "repo_link": "https://github.com/Creative-Difficulty/sentience-workers",
    "reported_seconds": 456328,
    "screenshot": "https://r2.hctg.gallery.karimeltaib.com/977.png",
    "status": "Approved on 2026-05-10",
    "submitted_at": "2026-05-10T20:59:46.070Z",
    "tags": [
        6
    ],
    "title": "Sentience",
    "total_seconds": null,
    "unread_notification_count": 0,
    "updated_at": "2026-05-10T21:04:33.965Z",
    "user_id": 490,
    "username": "oneonealex",
    "ysws": null
}