Build, run and scale AI agents like API and microservices - observable,auditable and identity-aware from day one.

Your machine's AI brain. One 20MB binary gives every tool, script, and cron job shared AI memory + 133 API routes. Desktop app, CLI, Telegram — all connected. Rust-powered.

A powerful, faster, scalable full-stack boilerplace for AI inference using Node.js, Python, Redis, and Docker

Create your backend with AI

Manifest Backend-as-a-Service Docs

A production-ready FastAPI with Pydantic AI backend service that provides OpenAI-powered chat completions with streaming support, designed for seamless integration with the Vercel AI SDK.

AI-powered backend for user requests, enabling seamless chatbot integration... Created at https://coslynx.com

Realtime voice AI backend with low-latency streaming using WebSockets

A clean, minimal RAG backend that demonstrates document ingestion, embedding, retrieval, and LLM-based answering with persistent local storage.

Módulo de inteligencia artificial para dabetai. Modelos de predicción de complicaciones diabéticas tipo 1 (retinopatía, nefropatía, neuropatía, pie diabético). Componente de backend para plataforma de monitoreo de la diabetes.

Backend-controlled AI feedback agent with memory, state guards, and safe tool execution.

AI-powered quotation intelligence microservice leveraging FastAPI, Pydantic schema enforcement, decimal-accurate pricing and tax computation, and bilingual (EN/AR) AI-driven email generation via an LLM abstraction layer. Features OpenAPI contracts, pytest-based CI validation, and Dockerized, cloud-native deployment for scalable CRM and sales automa

Async AI/HITL orchestration backend - FastAPI, hexagonal architecture, one time approval tokens, structured observability, and Telegram webhook integration. Built for secure, enterprise ready automation workflows.

FastAPI-based stateless inference service untuk serving multiple AI models with input video.

Enterprise RAG backend for knowledge retrieval using FastAPI, PostgreSQL/pgvector, AWS S3, and AWS Bedrock.

AI-Powered Home Surveillance System

Production-style AI backend platform built with FastAPI, PostgreSQL, RAG, and MCP. Demonstrates real-world AI platform engineering: typed APIs, RBAC, vector search, dockerized infra, and full pytest coverage.

Production-ready Retrieval-Augmented Generation (RAG) backend system with Pinecone, Groq, Docker & MLflow observability.

Node.js backend API for a video transcription prompt application. Built with Fastify, Prisma, Zod, and TypeScript. Supports MP3 video uploads, custom prompts with {transcription} placeholders, and streaming AI completions (e.g., via OpenAI) for interactive video content querying and generation.

Production-pattern backend service integrating LLM inference and vector similarity search with FastAPI, OpenAI, and ChromaDB