AI agent run Windows but with AWS Access sep 2026

@denisSolution

Hi, I'm Denis, a full-stack developer experienced in AI agents, RAG systems, and FastAPI applications with secure AWS integration. I understand you need an educational AI agent for Windows with AWS-hosted PostgreSQL, hybrid RAG, chat memory, and comprehensive security testing. You want to learn the entire system end-to-end and reproduce it independently, with full documentation through screenshots, videos, and modifiable code. I've built similar systems with FastAPI backends, Python frontends, conversation memory, and hybrid retrieval with configurable chunking. I'm experienced in Windows deployment with secure AWS access and credential handling. My approach: 1. Requirements clarification and architecture design 2. Build core application with login, chat, uploads, and admin controls 3. Implement hybrid RAG with configurable chunking and persistent memory 4. Deploy to Windows with secure AWS access and PostgreSQL/pgvector setup 5. Comprehensive security testing using Playwright 6. Create DOCX manual with screenshots and matching videos 7. Demonstrate clean rebuild from your own AWS account Key technical consideration: balancing RAG flexibility with clarity through configurable chunking, multiple retrieval modes, and detailed ranking inspection. Main risks: unclear RAG behavior specifics and AWS cost estimation. I'll validate assumptions early and provide detailed cost breakdowns with test runs. I'm ready to start immediately. Let's discuss the details. Thanks, Denis.

@extreamcode

Hi ClientFirstName, You’ve got a very specific hidden goal here: not just an AI agent, but a reproducible, teachable Windows-to-AWS setup with memory, hybrid RAG, and proof it survives real security and multi-user pressure. I’ve built Python and FastAPI systems with database-backed workflows, auth, admin controls, uploads, and AWS-connected deployments, so I can structure this as a clean educational build instead of a black box. I’ll create the example agent, wire the FastAPI app, memory, retrieval modes, admin controls, PostgreSQL/pgvector, and the Windows deployment path with secure AWS access using the browser-based console. I’ve shared an initial estimate based on your description, and once we go over a few technical or functional details, I’ll confirm the exact cost and delivery schedule. Could you confirm whether you want OCR excluded for the first version, and whether the Windows host will be a dedicated machine or shared workstation? Also, should the 25-user acceptance test prioritize concurrency under load, or longer multi-turn dialogue depth? Looking forward to your reply so we can finalize the exact plan. Best regards, Asad

@PINESCRIPTERR

Hi, I can build the Python/FastAPI AI agent with hybrid RAG, persistent memory, PostgreSQL/pgvector, secure AWS connectivity, Windows deployment, and comprehensive security testing using Playwright, HTTP tests, and the specified Patchright repository. I’ll also deliver matching source code, DOCX documentation, videos, tests, and reproducible deployment/recovery steps so the project can be rebuilt independently. 1. Which LLM and embedding provider do you prefer? 2. EC2 or Lightsail for the PostgreSQL/pgvector database? 3. Do you already have target latency/error thresholds for the 25-user load test? Regards, Houssame

@kumarashish99

As an AWS-certified professional, I bring over 5 years of specialized experience in backend development, DevOps engineering, and Kubernetes orchestration to the table—skillsets that are indispensable for your AI agent Windows project. Having built scalable, secure, and cost-efficient cloud infrastructures integrated with AI/ML elements before, I am uniquely positioned to help you with your goal of creating and running a Python/FastAPI AI agent on Windows with AWS Access. In terms of development, I am well-versed in Python and have extensive hands-on experience building serverless and containerized applications. With a strong background in DevOps, I can also implement infrastructure-as-code tools such as Terraform to expedite the setup and configuration process. My previous engagements with AWS platforms like EKS, GKE, and AKS ensures that I know my way around the AWS Console/Portal—an added advantage for this project. every stage of the project. My aim is not only to deliver an outstanding product but also enable you to understand every aspect of it independently ongoing from this project. chosen Promtheus New MongDB New RelicNew Relic

@mickeysz

Hello, The key constraint is the network path. The app stays on Windows, with no inbound port-forwarding or third-party tunnel. Remote users connect over HTTPS through AWS, with SSM preferred. The clean approach is an outbound reverse tunnel: the Windows host runs SSM Agent and connects outward to a small always-on EC2/Lightsail instance exposing the public HTTPS endpoint behind CloudFront and WAF. Nothing inbound is exposed on Windows, and the app remains on-prem. This AWS infrastructure is a real recurring cost and should be itemised. For hybrid RAG, pgvector provides dense similarity, while PostgreSQL full-text search (tsvector/ts_rank) provides the lexical side. It is not BM25, so I would not call it that. Hybrid retrieval combines both using reciprocal rank fusion. PostgreSQL stores vectors, keyword indexes, conversations, memory, and settings. Chunking and retrieval weights are admin-configurable. Reindexing is versioned, so changes trigger controlled rebuilds and prevent incompatible embeddings or partial indexes going live. Deletion removes active retrieval entries, with originals and backups handled separately. Agent actions are enforced server-side, so prompt injection in a document cannot exceed the allowed action set. One question: can the Windows host run SSM Agent and make outbound connections, and is a small always-on EC2/Lightsail instance acceptable as a recurring cost? That determines the topology.

@lazarr4

Hello, I can build a secure Windows-hosted AI agent with AWS access, using FastAPI and Python Playwright for testing, while keeping costs low and security tight. I built a similar system for a client—used PostgreSQL on Lightsail with FastAPI and Playwright tests, reducing deployment time by 30% compared to Docker. I’d approach this by running the agent directly on Windows with AWS access via SSM, using PostgreSQL for chat memory and RAG storage. The biggest improvement will come from isolating AWS access through IAM roles and testing every security layer with Playwright. Proper setup will reduce security risks and ensure reliable performance under load. I’ll provide a detailed DOCX manual and videos to teach you how to maintain and extend it. Once the project is complete, I’ll review the security tests and deployment with you to confirm everything is working as expected. If this sounds helpful, I’d be happy to discuss the details further. Thanks, Lazar.

@CorporateRD

Hey, I reviewed the complete scope and the supplied `patchright-enhanced` repository. You need more than an AI chatbot: a reproducible Python AI Agent with FastAPI, hybrid RAG, persistent memory, PostgreSQL/pgvector, AWS connectivity, security testing, load testing, monitoring, and complete teaching material. I will build the application in Python with FastAPI and a Python based UI, covering document ingestion, dense/keyword/hybrid retrieval, configurable chunking, reindexing, source citations, private/shared documents, user memory, admin controls, and versioned retrieval data. The main app will remain on Windows with secure AWS connectivity and a low cost PostgreSQL/pgvector database. I will document setup, recovery, backups, costs, and network architecture. For security testing, I will use the supplied repository with Python Playwright and HTTP tests, covering auth, access control, isolation, uploads, injection, and quotas. I will provide the editable Playwright/pytest suite, load tests, reports, source code, recovery scripts, screenshot based DOCX manual, and matching videos for independent reproduction. - Which LLM provider/model do you prefer, or should I select a cost effective option? - Do you already have the Windows hardware specification for the 25 concurrent user test? Hop on a call or hit the chat to discuss the architecture, security, testing, AWS setup, and acceptance criteria. Thanks, TF | Royal Designs

@shakilan4

Hi, I can build and document your Python/FastAPI AI Agent Tutor with hybrid RAG, persistent memory, AWS database, Windows deployment, security testing, Playwright automation, DOCX manual, and matching videos. I’ve read your project details carefully and will use the AWS Console only, with reproducible setup, secure access, versioned reindexing, monitoring, recovery, and real-model/browser tests. I’ll deliver a fully self-sufficient repository, screenshot-based DOCX, narrated videos, security/performance reports, and clean handover so you can rebuild and modify the project independently without my accounts or secrets. Ready to start immediately. Best regards, Shakila Naz

@melopgabriel

Hi, this project is clearly about building and teaching a secure Windows-based AI agent with AWS-backed access, memory, and hybrid RAG. I’ve built Python/FastAPI systems with role-based access, chat history, retrieval pipelines, and test suites using Playwright and direct HTTP checks. I can keep everything Python-led, with a compatible browser UI, pinned dependencies, and secret-free config templates. My approach would be to build a document assistant example end to end, then add memory, hybrid retrieval, admin controls, permissions, and the Windows/AWS networking path using the browser-based AWS Console only. I would also structure the tests, screenshots, reports, and DOCX manual so you can reproduce the setup independently. I can help you reach a working first milestone quickly and then harden it with security and recovery testing. Let’s discuss the exact scope and handover format. Best regards, Gabriel

@yasinu8

As an experienced and award-winning full-stack web developer, I have a proven track record in implementing and maintaining complex web applications like the one you're seeking. Especially given my skills in Python, FastAPI, and AWS, I believe I am perfectly suited to take on this educational project and teach you how to independently reproduce the AI agent you desire. Throughout our collaboration, I will prioritize your understanding by providing readable code, detailed documentation, and comprehensive videos which equip you to progress without dependence on specific formats. In handling the Python application, front-end, and admin tasks associated with your project, there is no doubt that my extensive knowledge would be crucial in ensuring a successful delivery. With my experience in storage management and deep understanding of managing, interacting with APIs and integrating third-party services in Python, your application will be airtight in security. Additionally, my prowess in generating relevant and well-structured logs will guarantee efficient debugging processes. Lastly yet importantly, deploying your application on a low-cost AWS database is one of my strong suits. My intimate familiarity with PostgreSQL with pgvector, full-text search systems as well as migration techniques will enable me set up a top-notch yet cost-efficient hosting system for your project.

@rakhshanz3

With almost two decades of experience in the field of computer science and artificial intelligence, I bring to the table the kind of expertise and comprehensive knowledge that your project demands. From my early days as a lecturer to my recent role as a Chief Technology Officer, I have always thrived on teaching and problem-solving alike. Specifically, I specialize in deploying Python-driven applications with FastAPI backends, making it easy for you to understand and reproduce the entire process from inception till final implementation. The fact that I have already built AI-powered voice and chat agents including solutions for document processing and data extraction, consistently leveraging on machine learning - aligns well with the requirements of your project. Moreover, given your need for secure AWS access at a low cost with strict compliance to PostgreSQL database, I'm well-versed in AWS setup and configuration and an expert in PostgreSQL management. My past experience with various multinational software companies has given me an in-depth working knowledge of AWS services and their best usage tactics. In summary, as an accomplished AI specialist with not just academic credentials but also proven professional achievements, I can ensure that your request for a DIY educational project is smoothly fulfilled. You can count on me to deliver not just code but clear reporting documentation inclusive of video tutorials or

@sofiaglzwork

I understood your project clearly and I’m excited to build this educational AI agent end to end. I have experience delivering FastAPI systems with secure login, chat history, memory, hybrid RAG, and Python Playwright testing. You need: 1) Windows-hosted app with AWS-secured access, 2) memory across sessions with inspect/delete controls, 3) hybrid retrieval with admin-tunable chunking, permissions, reindexing, and reports. My plan is to build a document knowledge assistant with FastAPI, Python UI, PostgreSQL + pgvector, FTS, and secure AWS connectivity using browser-based console setup. I’ll implement bounded agent actions, role-based access, encrypted secrets, tests, screenshots, DOCX manual, and demo videos. For implementation, I’ll need your preferred AWS region, expected document volume, and any branding details. I’ll work in milestones, provide progress demos, and target a clean reproducible handover. Sofia Belem

@sara0346

Hello, I reviewed your project for building and teaching a Python/FastAPI AI agent with hybrid RAG, persistent memory, AWS connectivity and security testing. I will build the application directly on Windows with PostgreSQL and pgvector, keeping AWS limited to the required low-cost database and supporting services. I will implement authentication, admin controls, document ingestion, configurable dense and keyword retrieval, memory, secure agent actions, Playwright tests, monitoring, recovery and performance testing. I will also deliver the complete source, security evidence, editable DOCX manual and matching videos so the project can be reproduced independently without my accounts. Should the first prototype use a document knowledge assistant as the example agent? Best Regards, Sara

@miguelcen64

I can build this end to end: a Python/FastAPI AI agent tutor with hybrid RAG, memory across sessions, secure AWS setup, Windows deployment, and reproducible testing. I’ll keep the implementation educational, self-sufficient, and fully documented, with source code, screenshots, DOCX, and videos aligned to the same steps. I’ll also wire in the required Artificial Intelligence, Python, Amazon stack in a way that supports browser-based AWS Console setup, secure access, and testable workflows. Which AWS region and Windows version should I target for the documented setup and testing? I’ll structure the project so you can reproduce it independently, with pinned dependencies, secret-free templates, admin controls, retrieval tracing, and security testing against the authorized repository you specified. My approach will emphasize clarity, repeatability, and handover quality. Best, Miguel

@ankits0192

Hi — I read your full brief. The core challenge isn't building an AI agent; it's making it reproducible by you: Windows-hosted FastAPI + Python UI, hybrid RAG with selectable retrieval and admin-tunable chunking, persistent per-user memory across restarts, low-cost AWS Postgres/pgvector over HTTPS with no tunnels, plus Playwright concurrency tests, a screenshot DOCX manual and matching MP4s. That scope is a multi-week build, not a 4-day job — and a $10–30 CAD fixed price won't cover it. I'd rather be straight with you than take it and underdeliver. If you're open to it, I can scope it properly and quote realistically — and deliver the DOCX/video teaching material as the strongest part. Three questions: 1. Is the budget flexible, or is $30 CAD the hard ceiling? 2. Is 4 days a real deadline or a placeholder? 3. Must all 50-user tests be real-model, or is 5–10 live plus mocked higher loads acceptable? Happy to discuss. Ankit

@engruhulajom

Hello Dear, I’m Md Ruhul Ajom, an experienced Python & AI Developer with 10+ years of experience building FastAPI applications, RAG systems, AWS deployments, security testing, and automation. I understand you need an educational Windows-based AI agent with hybrid RAG, persistent memory, PostgreSQL/pgvector, AWS access, security validation, Playwright testing, monitoring, recovery, and complete reproducible documentation/videos. I am skilled in Python, FastAPI, RAG, PostgreSQL, pgvector, embeddings, Playwright, pytest, AWS, CloudWatch, authentication, security testing, and Windows deployment. I have a few questions: 1. Which LLM provider should be used? 2. Is OCR required? 3. Which Windows version will you use? I can build the project with reproducible setup, security evidence, performance tests, DOCX instructions, videos, and complete handover. Looking forward to hearing from you. Best regards, Md Ruhul Ajom

@muhammadsaad90

I’ll build and deploy the complete Python/FastAPI AI Agent Tutor on Windows with secure AWS access, low-cost PostgreSQL/pgvector hybrid RAG, persistent memory, admin/user controls, security testing, and Python Playwright load tests—using the AWS Console only, with no Terraform or Docker. In 4 days, I’ll deliver the fully reproducible source code, clean-Windows/AWS deployment, security & performance reports, editable screenshot-based DOCX manual, and matching step-by-step MP4 videos, so you can independently rebuild, modify, test, and operate the entire project without my accounts or assistance.

@Cruz1217

Hi, running a Python agent on Windows with secure AWS access requires careful network path planning to avoid exposing credentials. I would set up a direct HTTPS tunnel from the Windows host to AWS, using SSM for management, and build the FastAPI backend to handle the RAG logic and user sessions locally. This keeps the main application on Windows while securely connecting to the database. A common mistake is treating the AWS database as a public resource, which often leads to security issues. I will implement strict origin restrictions and least-privilege access so the Windows host can only talk to the database. You will get a working system with a clear manual and videos that show exactly how to reproduce the setup.

@felixivanr

Hi, The core challenge is building a real, teachable AI agent system on Windows with secure AWS access—not a demo—while proving it's secure and reproducible without embedded credentials. The hybrid RAG with configurable chunking, versioned reindexing, and admin inspection of retrieval scores is the technical centerpiece. I'd use FastAPI with PostgreSQL+pgvector on Lightsail, storing conversations, embeddings, memory, and settings in one database. The Windows app runs via Systems Manager Session Manager for HTTPS access (no tunnels). For RAG, I'll implement BM25 lexical search alongside dense embeddings with configurable fusion weights. Memory isolation separates chat history, persistent user memory, and RAG context—users can inspect and delete each independently. Testing uses Python Playwright against the real frontend with concurrent users and repeatable YAML dialogues. Security testing follows your patchright-enhanced repository: I'll pin the revision, run Playwright and direct HTTP attacks, and report browser detection separately from auth failures. I'll test injection, path traversal, XSS, CSRF, malicious prompts, and quota enforcement—no WAF-only acceptance. All source, config, and secrets stay off AWS; you deploy using only the DOCX manual, videos, and scripts I provide. I'll deliver the codebase, DOCX with numbered screenshots and copyable commands, MP4 videos for each step, load-test results with latency/throughput/costs, and a clean rebuild demo on fresh Windows