
Closed
Posted
Paid on delivery
Summary Job Title: AWS AgentCore phone voice agent coded in Python 5 days to do do not use Amazon Nova Provide minimum delays between question – answer – question for conversation dialogs Use or [login to view URL] Posted on: Sep 22, 2026 with AI agents outside Connect Customer, over the open agent-to-agent (A2A) protocol through text or bidirectional voice AWS expanded the protocol to support voice streaming Customer phone ↕ Amazon Connect Customer ↕ A2A voice streaming ↕ AWS AgentCore (Python) ├── Speech-to-text ├── Bedrock LLM └── Text-to-speech Or just idea Phone provider: Twilio can't connect directly to AgentCore, because its stream address doesn't allow the login signature AgentCore needs. Vonage can, and AWS and Vonage have a sample that works with real phone calls, so I planned to use Vonage. Voice code: the Pipecat library (version 1.12) has everything we need in Python. It handles Vonage audio, Transcribe, Polly and OpenAI, and it can change the system prompt during a call, which is what your situation-based prompt needs. OpenAI on AWS: GPT-5.6 Luna can be called through Bedrock from Canada (Central). Requests go to Canada or US regions. Deploy tool: AWS has replaced the old Python agentcore command with a new one installed through npm (@aws/agentcore). [login to view URL] Job Type: Fixed-Price Very good natural voices for LLM use low cost fast models from openai or gemini or make suggestion **** No Terraform or similar techniques, only simple browser aws platform development *** use AWS from local windows Project Overview **** Develop Voice Real Estate agent (sell homes) **** For testing provide realistic properties descriptions from [login to view URL] or [login to view URL] or [login to view URL] etc. convert property web page to text extract Neighbourhood description and statistics Like example https://www.realtor.ca/real-estate/30180444/829-55-stewart-street-toronto-waterfront-communities-waterfront-communities-c1#view=stats https://www.realtor.ca/real-estate/30180444/829-55-stewart-street-toronto-waterfront-communities-waterfront-communities-c1#view=neighbourhood [login to view URL] I am looking for an experienced AWS and Python developer to build a scalable Voice AI Agent architecture using AWS Bedrock AgentCore. This is fundamentally an educational project: your goal is to build a fully working example system, document it exhaustively, and teach me how to reproduce, modify, and deploy it completely independently from scratch. looking for a one tenant Voice AI platform where agents can be created and managed, with Canadian phone number, isolated skill files, call memory/logs, and admin management, built on AWS Bedrock AgentCore + Python, along with complete documentation, videos, testing, and handover. You may use existing open-source GitHub projects as a foundation, provided the final deliverable meets all my requirements and is fully documented. agent answer questions about properties and persuade user to buy and book viewings intergrade some useful booking service and send message by email and sms text to both user and realtor full log such call is transcribed and session logs System Architecture & Core Examples We need a one Canadian phone voice agent system acting as customer service caller phone number ID identification email interface for admin to provided gmail address : add or remove files for skills; send by email logs of conversations , block phone numbers for incoming call The Use-Case Examples: Think of these as sales agents for stores or services , where store has many unique products, for example : Or Case 1: Food Catering Or case 2: Cleaning Services Or case 3: Refrigerator Repair Or case 4 : Real Estate agent (sell homes) Expected Interaction Flow: At the beginning of the call, the agent asks the user what they want to talk about (essentially asking which file to add to system prompt ). Example A (Case 3): A user calls for Store 3 and says, "I have Refrigerator abc123." The agent must dynamically retrieve the specific .txt skill file for "Refrigerator abc123" and use this skill for the remainder of the session. Example B (Case 2): A user calls for Store 2 and says, "I need to clean a sofa." The agent chooses the specific .txt skill file on how to clean a sofa and uses it for the session. Example C (Case 3) : user wants to talk about specific home for sale . Find skills file with specific address what user tells you . Each .txt file has header with description what is this file about , then agent find which file header matches the beset to user request , if bad matching happen then during conversation with user it should be found correct match , etc . Usual skills router implementation Also if user wants then conversations returns to state to choose what property to discuss **** Also search of matching properties by search criteria for example all properties from your portfolio with one bedroom , test with 10 properties in portfolio *** Technical Features & Requirements Voice Integration: Phone call voice chat using a Canadian phone number (+1). One phone number serves exactly one agent/store. Strict Data Isolation: agent has its own dedicated group of .txt knowledge files. Session Memory: The agent must retain context during the call (e.g., if the caller says "My name is Peter," the agent calls him Peter for the whole session). There must be no memory between different sessions. Logging: Detailed logs must be generated and saved for every individual call/session. Technical Stack Language: Python AI Framework: AWS Bedrock AgentCore Telephony: use Amazon Connect - low cost AWS services compatible with AgentCore , I am open to discuss other options like Chime SDK, or Twilio Phone call with interruption option Infrastructure: AWS Strict Development Constraints Your Own Environment and development resources : You must develop, record, and test this entirely in your own authorized AWS account. I will not provide my AWS account or LLM credentials (openai or gemini ) or etc resources Zero Dependencies: My later deployment must not depend on your accounts, resources, or API keys. Use openai or gemini for LLM (I am open for suggestions for other good LLM providers ) Deliverables (Required for Payment) This fixed-price project covers the application, deployment, security, tests, source code, reporting, manuals, videos, and a handover session. 1. Editable Word (.docx) Manual: Must start from a totally clean environment. Every actionable setup/configuration step needs a real, readable screenshot. Numbered instructions with copyable commands/settings. Explain the purpose of the step, the expected result, and how to verify it. Clearly distinguish between Windows and Server commands and explain any placeholders (e.g., [YOUR_BUCKET_NAME]). Must cover: Prerequisites, code structure, AgentCore setup, memory/RAG routing, telephony settings, database/logs, AWS IAM/security, tests, monitoring, recovery, updates, cost breakdown, and complete environment cleanup. Include exact software/library versions and architecture details. Do not include real secrets/passwords in the doc. Include a section on common errors and fixes. 2. Narrated MP4 Videos: Follow the exact section/step numbers of the .docx manual. Include a timestamp index. Show the actual, complete setup from scratch (no skipped prerequisites or undocumented pre-configuration). Demonstrate application features, parallel-dialogue tests, and a security assessment showing the strict .txt file isolation. Note: Videos, screenshots, and code must match perfectly. 3. Handover & Teaching: A session (or clear video guidance) teaching me how to modify both the Python application codebase and the Python test suite. Acceptance Criteria The project will be considered successful and complete only when I am able to reproduce the entire project by myself, from scratch, in my own AWS account, relying entirely on your .docx manual and videos without needing to ask you for missing steps. Very good natural voices Clear instructions how to delete all created resource – not to pay after project done
Project ID: 40757543
79 proposals
Remote project
Active 1 min ago
Set your budget and timeframe
Get paid for your work
Outline your proposal
It's free to sign up and bid on jobs
79 freelancers are bidding on average $52 CAD for this job

Hello I checked your project, and the main point is to build a reproducible Python voice real-estate agent on AWS, with Canadian phone support, dynamic .txt skill routing, session memory, call transcription, booking, notifications, and complete teaching materials. I have experience building Python AI applications with document processing, LLM workflows, data extraction, and session-based logic. I can also use the Pipecat approach with Amazon Transcribe, Polly, Bedrock, and Vonage or Amazon Connect, depending on which voice path is most reliable for AgentCore. I would first create a clean one-tenant architecture, then implement property text extraction, skill-file matching, conversation state, interruption handling, and isolated call sessions. Call logs can be stored per session, with booking and email/SMS actions added through AWS services, followed by testing with ten sample properties and clear deletion procedures. The most important part will be matching the .docx manual, screenshots, narrated videos, and deployed code so you can reproduce everything from a Windows environment without my AWS account or credentials. Could we clarify whether you prefer Amazon Connect or Vonage for the first working version, and which booking service should receive viewing requests? I’d also like to confirm whether OpenAI or Gemini must be used, or if a Bedrock-native model is acceptable for lower operating cost. Best regards
$20 CAD in 5 days
7.6
7.6

Hi, I'm Denis, a full-stack developer experienced in building voice AI systems that connect phone calls to LLMs with dynamic skill routing. You need a working voice agent platform on AWS that handles real phone calls, routes callers to the right knowledge files based on intent, logs everything, and runs in your AWS account. The agent must work for different use cases—real estate, repair services, catering—by swapping skill files without redeploying. You want to learn how to build and modify this yourself, not depend on me after handover. I'll start by setting up AgentCore with Vonage for phone integration and Pipecat in Python to handle voice streaming, transcription, and LLM calls. Next, I'll build the skill router that matches caller requests to the right .txt file while maintaining session context. Then I'll add call logging, SMS/email notifications, and admin controls for managing skills and blocking numbers. Finally, I'll test the full flow and document every step with screenshots and videos. The core challenge is the skill router—it must reliably match user intent to the correct file and switch if needed. I'll use prompt engineering and confidence scoring to handle this. Documentation gaps in AgentCore and Vonage could slow setup, so I'll build and test everything in my account first, then create detailed step-by-step guides you can follow. I can start right away. Let's connect and discuss the details. Thanks, Denis.
$30 CAD in 1 day
6.0
6.0

Hi there, I'm excited about the opportunity to work on developing a scalable Voice AI Agent architecture using AWS Bedrock AgentCore. Your project aligns perfectly with my expertise in Python, AWS, and AI-driven voice solutions. I understand the objective of creating a fully functioning educational system where I will not only deliver a robust voice agent but also equip you with comprehensive documentation and guidance for future independence. With over five years of experience in AWS and AI, I have successfully implemented similar projects, ensuring seamless integration with telephony services like Amazon Connect and Vonage. My approach will involve leveraging open-source libraries such as Pipecat for handling audio, Transcribe, Polly, and OpenAI, ensuring smooth voice interaction and dynamic skill retrieval during calls. I'll ensure strict data isolation and session memory, maintaining context throughout conversations. I'll provide thorough documentation and video tutorials, making sure you can replicate and modify the system with ease. My approach focuses on building a one-tenant platform with features like call logs, skill file management, and Canadian phone number integration, tailored to real estate or other service industries. I'm eager to contribute my skills to bring your vision to life and empower you with the knowledge to manage your Voice AI platform independently. Best Regards,
$10 CAD in 2 days
5.4
5.4

Hi. To build this reliably within five days, I would create a Python voice-agent proof of concept using Amazon Connect or Vonage, Pipecat, AWS Transcribe, Polly, Bedrock, and AgentCore, with interruptible bidirectional audio and per-call session memory. The architecture will use a clean skill-router layer for isolated property files, structured retrieval, property search, viewing bookings, SMS/email notifications, caller blocking, and complete transcription and session logging. I will keep deployment browser-based through AWS services, avoid Terraform, and provide a small test portfolio of realistic Canadian property records with neighbourhood and statistics content. The implementation will include IAM boundaries, tenant/file isolation tests, cost-conscious model selection, monitoring, cleanup instructions, and reproducible Windows setup steps. As a Senior AI Engineer, I have mastered Python, AWS Bedrock, AgentCore, Amazon Connect, Pipecat, voice pipelines, and RAG routing, with strong experience in conversational AI and cloud automation projects. I will document the code, tests, architecture, screenshots, deployment, security checks, and handover process in an editable manual, supported by narrated walkthrough videos. I am sure I can deliver high-quality results within the requested five-day timeline for a focused working example and clearly identify any production extensions. Let’s get in touch and discuss more. Thanks.
$30 CAD in 5 days
5.1
5.1

Hello,, I hope you're having a great day. After reviewing AWS AgentCore phone voice agent coded in Python, I'm confident that I can deliver the results you need on time and within budget. I have a strong portfolio showcasing my work in Python, Amazon Web Services, Artificial Intelligence, Voice Agents and Python, Amazon Web Services, Artificial Intelligence. Please check my profile for examples relevant to AWS AgentCore phone voice agent coded in Python. Looking forward to bringing my Python, Amazon Web Services, Artificial Intelligence, Voice Agents expertise to AWS AgentCore phone voice agent coded in Python. I'm ready when you are! Regards, Manisha
$20 CAD in 1 day
4.3
4.3

I can help you build the AWS Bedrock AgentCore based Voice AI platform using Python with a focus on a reliable architecture, clean documentation and complete handover. I have experience building AI automation systems, LLM workflows, API integrations, backend services and cloud deployments using Python, AWS, RAG style knowledge retrieval and conversational systems. For this project, I would design the system with isolated skill files, dynamic knowledge routing, session based memory, call transcription, logs, email and SMS notifications, property search capability and admin management features. I would use a scalable approach with AWS services, Python voice pipeline components and secure storage so each tenant or business agent has completely separated knowledge and conversation history. I understand the importance of proving the system can be reproduced from a clean AWS account, so I will provide structured source code, setup documentation, deployment steps, testing procedures and resource cleanup instructions.
$100 CAD in 4 days
4.3
4.3

noticed your requirement and here is my understanding - Building a scalable, educational Canadian Real Estate Voice AI Agent using AWS Bedrock AgentCore, Python, Vonage telephony, and dynamic .txt skill routing with full isolation and booking integration. Stack: Python, AWS Bedrock AgentCore, Vonage, Pipecat, OpenAI/Bedrock LLM, AWS Console. I've gone through your requirement for the AWS AgentCore real estate voice agent. I can build this with your preferred stack - Python, AWS Bedrock AgentCore, Vonage, and Pipecat - with clean, scalable codebase within 7 days. Milestone 1: Core Voice Architecture & Dynamic Skill Router Implement Vonage-to-AgentCore voice streaming, Pipecat state management, dynamic .txt real estate/neighbourhood portfolio routing, and session memory handling on AWS. Milestone 2: Booking, Notifications, Logging & Handover Integrate SMS/Email notifications for user and realtor, comprehensive session call logging, end-to-end testing, exhaustive .docx manual with screenshots, narrated MP4 setup videos, and cleanup scripts. Let's review target timelines and map core milestones. I may have few questions before suggesting full architecture. Lets hit CHAT and discuss now ! Gaurav P. 15 yrs experience in debugging systems :)
$99 CAD in 3 days
4.1
4.1

Hi, I will start by building the call flow around Amazon Connect → voice streaming → Python/AgentCore → STT/LLM/TTS, with interruption handling and minimal question-to-answer latency. I’ll implement per-agent .txt skill isolation, semantic skill routing, session-only memory, property search, call transcription/logging, and booking/email/SMS actions without carrying memory between calls. I’ll build the admin flow for managing skill files, blocked numbers, agents, and conversation logs, then test the real Canadian phone workflow with a 10-property dataset and parallel call scenarios. I’ll deploy everything in an isolated AWS environment, document every dependency/configuration, and provide reproducible setup, security tests, videos, source code, and cleanup instructions so your own AWS account has no dependency on mine. For the skill router, should matching be embedding/RAG-based with the selected .txt skill injected into the session context, or do you want a deterministic header/category match first with LLM fallback? I can design the routing and isolation around that decision and start with the end-to-end call prototype.
$10 CAD in 1 day
4.1
4.1

Hello, I can build the voice agent using AWS AgentCore with Python, handling real-time voice streaming and dynamic skill routing while keeping costs low and ensuring strict data isolation. I built a similar system for a real estate client—used Pipecat 1.12 with Vonage for voice streaming and Bedrock for LLM calls, reducing deployment time by 40% by avoiding Terraform. I’d approach this by using Amazon Connect for telephony, AgentCore for the voice agent core, and a custom .txt file router to switch skills mid-call. The biggest improvement will come from isolating each tenant’s knowledge files in S3 buckets. Proper setup will reduce call drop rates and improve agent responsiveness by 30% through optimized voice streaming. Once the project is complete, I’ll provide a detailed cleanup guide to ensure no AWS resources are left running after the project ends. If this sounds helpful, I’d be happy to discuss the technical details further. Thanks, Lazar.
$10 CAD in 1 day
3.9
3.9

With more than 8 years of comprehensive experience in full-stack development, I bring a unique blend of technical expertise and innovative problem-solving to this project. My résumé spans across multiple domains, building reliable and scalable solutions using languages like Python and platforms like AWS, which are central to your project. Having effectively integrated AI and ML models in various applications, like time-series forecasting and document intelligence, I can confidently design a robust Voice AI Agent around your requirements. Over my career, I've honed my cloud computing skills with AWS as well as my proficiency in programming in Python – both integral to your project's success. In line with your preference for an AWS-based system developed solely on the browser, my prowess with these technologies will eliminate the need for Terraform or similar techniques while providing straightforward yet effective AWS platform improvements that will fit snuggly alongside your current architecture. Additionally, apart from technical competencies, I'm well-versed in remote collaboration, maintaining clear communication channels throughout the project lifecycle. Combining my strong problem-solving abilities and passion for creating value-driven products, I ensure an educational journey through this project that will arm you with the skills to manage and modify the Voice AI system without any reliance on outside expertise.
$20 CAD in 1 day
4.1
4.1

AgentCore with Python for the phone agent, right call since it handles the session and memory layer for you instead of rolling your own state machine. I would wire up the voice pipeline with Amazon Connect or Twilio for the telephony leg, Bedrock for the LLM calls. Can start today, working version in 3 days. These numbers are initial estimates from the post, happy to firm them up once we talk through the exact call flow. Want me to send a quick plan?
$30 CAD in 4 days
3.7
3.7

Hi there, this project on developing a scalable Voice AI Agent using AWS Bedrock and Python aligns well with my expertise. I can utilize the Pipecat library for effective voice interactions through Vonage, integrating features like real-time transcriptions and dynamic skill retrieval based on user prompts. For the expected interaction flow, I would ensure that the agent retrieves specific .txt skill files according to the user's requests, facilitating a smooth conversation related to real estate. To meet your documentation requirements, I can provide detailed manuals and instructional videos that guide you through setting up the environment entirely on your own AWS account. I'm confident in delivering a robust solution that allows for context retention during calls while ensuring strict data isolation for each agent's operations. Looking forward to your response. Regards, Ali
$30 CAD in 1 day
3.4
3.4

Hello, I'm based in Japan, blending strong technical skills on Python and AWS with Japanese discipline, punctuality, and attention to detail. I'll work efficiently to complete your project perfectly, and as quickly as possible. Expect clear communication, reliable delivery, and a deep commitment to quality. Let’s make it happen. Warmly Regards.
$30 CAD in 5 days
3.2
3.2

Hello, Please provide the exact AgentCore/AWS requirements you want validated first. I can build a focused Python proof of concept for the Canadian phone voice flow, AgentCore/Bedrock integration, low-latency speech pipeline, session-only memory, dynamic .txt skill routing, interruption handling, and call transcription/logging. I’ll keep the architecture independent of my AWS account and credentials, with your future deployment able to use its own resources and keys. I can also document the setup and testing clearly. The complete production platform, multi-property search, booking, admin tools, videos, full Word manual and exhaustive from-scratch deployment would need a separate budget and milestones beyond the posted CAD $30. Best Regards, Sara
$170 CAD in 1 day
3.4
3.4

With my extensive skills in Python and AWS, I bring an ideal combination of technical expertise and real-world experience to the table. As one of the architects behind intricate digital platforms and scalable web applications, I have designed and implemented solutions that can handle large-scale data flows and support the complex workflows characteristic of both AI-driven systems and telephone-based interactions. Drawing on this background, I am trained in transforming abstract requirements into tangible and effective digital experiences. Lastly, as an experienced professional who prioritizes collaboration and thorough documentation, I guarantee a fully-documented project that you can independently manage from scratch after handover. My track record includes not just delivering robust platforms like the ones used by DEXUtils or C4ADS Seamless Horizons but also producing training materials for seamless adoption. By choosing me for this project, you're choosing more than just a short-term development solution; you're investing in a long-term partner who will equip you with comprehensive knowledge for successful digital transformations.
$20 CAD in 7 days
3.1
3.1

I understand you need a reproducible educational voice agent on AWS Bedrock AgentCore: Canadian number, dynamic .txt skill routing, session-only memory, call logs, booking with SMS/email, and complete DOCX manual + narrated videos + handover. I can build and test everything in my own AWS account using only the browser console (no Terraform), so you can replicate it independently. My approach: Vonage (works with AgentCore) + Pipecat 1.12 in Python for bidirectional voice, AWS Transcribe/Polly, low-cost LLM (OpenAI/Gemini), dynamic skill router that selects the correct .txt file based on caller intent, session memory, full transcription/logging, and real-estate portfolio search (10 properties). I'll deliver a working prototype in 3 days, then complete documentation and videos by day 5. Quick question: which telephony do you prefer – Vonage (sample works) or Amazon Connect? And which LLM: OpenAI or Gemini? I have 5 years in Python/AWS and 2+ years in AI voice agents, with hands-on experience in Pipecat, Bedrock, and RAG routing. I'll provide the editable DOCX manual with screenshots, narrated MP4s, handover session, and cleanup instructions. Available to start immediately.
$30 CAD in 5 days
2.9
2.9

The Vonage and Pipecat stack is the right call for low-latency A2A voice streaming, avoiding the Nova limitations you mentioned. I recently built DocuMind AI using FastAPI and LlamaIndex, which gives me the exact RAG and Python backend expertise needed for your Bedrock agent. You can see that work at usmanahmed-dev.vercel.app. Are you planning to scrape real-time property data or use a static dataset for the initial test properties, and which specific Gemini or OpenAI model do you prefer for the LLM layer? My plan is to build the isolated skill files and deployment scripts using the new npm CLI, ensuring you have full control from your local Windows environment. I am ready to start immediately and deliver the complete educational documentation you require.
$30 CAD in 7 days
2.8
2.8

Hi there! I understand the main challenge is building a working AWS Python voice agent with reliable phone conversations, dynamic skill routing, strict tenant isolation, call memory, detailed logs, and complete from-scratch documentation. I have relevant experience with Python, AWS, AI integrations, voice agents, APIs, LLM workflows, session memory, and cloud deployments. I can work with AWS AgentCore, Bedrock, Amazon Connect, and suitable telephony or voice streaming components while keeping the architecture simple. I will build the Canadian phone flow with interruption support, speech-to-text, LLM and natural text-to-speech, then implement the .txt skill router so each call uses only the relevant knowledge files. I will add property search, booking, email/SMS notifications, transcripts, session logs, phone blocking, and admin file management. I will also test isolation with parallel conversations and document every setup step with matching screenshots, commands, videos, security checks, costs, and cleanup instructions. The work can be completed within your requested 5-day development window, with documentation and testing included in the final handover. check our work https://www.freelancer.com/u/ayesha86664 Do you already have a preferred Canadian telephony provider, or should I evaluate Amazon Connect versus Vonage for the lowest-cost reliable setup? Let me know if you’re interested & we can discuss it. Best Regards Ayesha
$20 CAD in 3 days
2.5
2.5

Dear Client, I can build your Python real-estate voice agent using AWS Bedrock AgentCore, with a Canadian phone number, natural two-way conversations, property search, viewing bookings, and SMS/email notifications. I understand the main goal is not just a working system, but one you can reproduce independently. At Bell Canada, I built production AI agents using Python, LangGraph, RAG, and LLM APIs, including dynamic knowledge retrieval, tool routing, session state, and automated testing. For your project, I'd implement isolated .txt skill files, address-based matching, searches across 10 properties, in-call context switching, and session-only memory. I'd evaluate Amazon Connect/A2A against Vonage for real-time audio, interruptions, latency, and Canadian phone support. The AWS setup would use the console and Windows tools without Terraform. I'd also implement transcription logs, phone blocking, admin controls, booking integration, and pytest tests covering knowledge isolation and concurrent calls. The five-day timeline is tight for the full documentation and video requirements, so I'd prioritize a working end-to-end system and agree on acceptance checkpoints. I'll provide source code, a screenshot-based Word manual, narrated walkthroughs, deployment instructions, cost estimates, and complete AWS cleanup steps. Best, Vadim
$10 CAD in 1 day
2.3
2.3

Hey, I can start on this pretty much right away. Using Vonage for A2A voice streaming is a smart choice since it can directly connect to AgentCore, while Twilio won't work due to the login signature issue. Plus, leveraging the Pipecat library to handle audio and dynamically change prompts during calls adds a unique edge that can enhance the voice agent's responsiveness. Here if you want to dig into the details. Thank you!
$24 CAD in 5 days
1.4
1.4

Downsview, Canada
Payment method verified
Member since Nov 29, 2017
$10-30 USD
$30-250 CAD
$10-30 USD
$10-30 CAD
$10-30 CAD
₹400-750 INR / hour
$250-750 USD
$30-250 USD
$30-250 USD
₹1500-12500 INR
$300-800 USD
$750-1500 AUD
₹12500-37500 INR
$30-250 CAD
$10-30 USD
₹600-1500 INR
€6-12 EUR / hour
€30-250 EUR
₹600-1500 INR
$250-500 USD
$30-250 USD
$20-9000 USD