DeepSeek V3

Top Builders

Explore the top contributors showcasing the highest number of app submissions within our community.

DeepSeek V3

General
Release date	2024
Author	DeepSeek
Website	DeepSeek Models
Repository	https://github.com/deepseek-ai
Type	MoE (Mixture of Experts) Language Models

The DeepSeek V3 model represents our most advanced AI architecture, designed for complex reasoning tasks and code generation. With enhanced context handling and improved instruction following, this model excels in technical applications and enterprise deployments.

Key Features

DeepSeek-V3: 671B parameters (37B activated per token), optimized for math, code, and multilingual tasks.
Code Generation: Supports 12+ programming languages
Advanced Reasoning: Chain-of-thought capabilities for multi-step problems
Enterprise-Grade Security: Built-in content filtering and compliance features
Speed: 3x faster generation than previous versions (60 TPS)
Open-Source: FP8/BF16 weights available on Hugging Face

Usefull Links

👉 Local Deployment Guide for DeepSeek V3 👉 Model Weights on Hugging Face 👉 API Documentation 👉 Deepseek V3 Paper 👉 Performance Highlights

Edit on GitHub

Deepseek DeepSeek V3 AI technology Hackathon projects

Discover innovative solutions crafted with Deepseek DeepSeek V3 AI technology, developed by our community members during our engaging hackathons.

Competitor Intelligence Firehose

Competitor Intelligence Firehose is an automated competitor monitoring system built for the Web Data UNLOCKED 2026 Hackathon. It monitors four major technology companies (Microsoft, Google, Amazon, Apple) by scraping their public websites and LinkedIn company pages. The project demonstrates real-world business value by calculating Return on Investment (ROI) for automated competitor research. Manual competitor monitoring typically takes 8 hours per week for a business analyst at $75/hour, costing companies $31,200 annually in labor costs. This solution automates that entire process using Bright Data's web infrastructure. The system is configured with a Bright Data API key and $250 in credits, ready to integrate Web Unlocker, Browser API, and Proxy Network products for production-scale deployment. Key features include: - Automatic competitor data collection - Success rate tracking and error handling - JSON report generation for easy integration - ROI calculation showing 5,206% return on investment The architecture is designed to scale from 4 competitors to 100+ by leveraging Bright Data's infrastructure for bypassing CAPTCHAs, rotating IP addresses, and rendering JavaScript-heavy pages. This reduces maintenance overhead from hours per week to zero while providing always-fresh competitor intelligence. Technical implementation uses Python 3 with the requests library, runs on Termux (Android), and is version-controlled on GitHub. The project successfully scrapes 4/4 competitor sites and generates comprehensive JSON reports ready for enterprise analytics pipelines.

Folio

Folio is a research agent for drug discovery. Give it a target protein, gene, or disease area, and it produces a structured briefing on the recent literature: top compound candidates ranked by evidence, mechanisms of action, key papers, and open questions, with every claim citing the underlying paper. The pipeline searches PubMed via E-utilities, retrieves the top abstracts, runs an extraction pass on each paper in parallel to pull compounds and mechanisms as structured JSON, then synthesises the extractions into the final briefing. Compound candidates are cross-referenced against ClinicalTrials.gov v2 so the briefing notes the highest clinical phase reached. Two Featherless models do distinct work: a small fast model handles the per-paper extraction fan-out, and a stronger model does the synthesis. Progress events stream over Server-Sent Events, so the UI updates one paper at a time as extractions finish. The source list stays visible alongside the briefing, with the full abstract and the extracted fields one click away, so a scientist can audit any cited claim back to the paper it came from. The web UI persists history per signed-in user via Supabase, exports server-side PDFs via WeasyPrint, and ships alongside a CLI for scripting and CI.

ChorusOps — Voice-Native Dealflow Orchestrator

ChorusOps is a voice-native dealflow orchestrator built for investors and enterprise teams who run deal discussions inside Discord voice channels. Instead of taking manual notes or switching to a CRM mid-call, ChorusOps listens, understands, and acts autonomously. The system captures Discord voice audio (Opus stereo 48kHz), downmixes it to mono, and streams it in real-time to Speechmatics' WebSocket API for highly accurate transcription with multi-speaker diarization — attributing every spoken sentence to the correct speaker automatically. Transcripts are routed to Gemini 2.5 Flash, which acts as a multi-step planning orchestrator. Using function calling, Gemini maintains a persistent deal state (deal name, stage, team notes, market context, funding ask) across the entire conversation via structured tool calls. When sufficient context is gathered, Gemini autonomously dispatches a DEEP_ANALYSIS job to a Featherless serverless inference worker running an open-source LLM. This worker produces a scored investment scorecard — including investment score, recommendation, strengths, and risks — which is automatically posted back to the Discord text channel and spoken aloud via Kokoro TTS. The bot supports barge-in interruption: if a user starts speaking while the bot is talking, TTS stops instantly. Multi-guild isolation ensures the system runs across multiple Discord servers simultaneously. Slash commands (/join, /say, /status, /tts, /voice) provide a full text fallback interface. ChorusOps targets the Agentic Workflows track: the agent plans its own steps, calls external tools, manages async multi-step tasks over time, and posts results without any human intervention — from first spoken word to final scored deal. Tech stack: Discord.js, Speechmatics RT API, Gemini 2.5 Flash, Featherless LLM inference, Kokoro TTS, Express, TypeScript.

VIGILO - Global AI Compliance Enforcement Engine

VIGILO is a full-stack autonomous AI compliance enforcement engine built for the EU AI Act. Unlike chatbots that wait for user instructions, VIGILO operates as a true agent: upload one document or code file describing your AI system, and VIGILO takes over completely. Its multi-agent pipeline — powered by Google Gemini 2.0 Flash for analysis, Featherless Llama 3.1 70B for obligation scoring, and Speechmatics for voice input — autonomously reads the submission, classifies the EU AI Act risk tier (Unacceptable/High/Limited/Minimal), maps it to applicable Annex III categories, scores compliance across Articles 9-15, and generates a detailed remediation report. Every step streams live to the UI via WebSockets. VIGILO covers both document compliance (policy docs, system descriptions) and technical compliance (code scanning for prohibited patterns, missing human oversight, lack of audit trails). Deployed on Vultr, orchestrated visually via n8n, and open-source under MIT. Target users: legal teams, CTOs, and AI product managers at companies deploying AI systems in the EU who face real compliance obligations coming into force in 2025-2026.

Kraken xStocks AI Trading Agent

Kraken xStocks AI Trading Agent is an autonomous trading system designed to analyze, execute, and optimize financial decisions in real time. Built on MeDo’s no‑code AI infrastructure, it integrates Kraken’s API for live market data, enabling predictive analytics, backtesting, and automated trade execution. The agent leverages reinforcement learning and sentiment analysis to adapt to market volatility while maintaining risk‑adjusted performance. Its dashboard visualizes portfolio metrics, trade history, and performance trends, empowering users to monitor and refine strategies dynamically. The project demonstrates scalable AI‑driven finance automation with transparent analytics and modular workflow design.