AI Receptionist Web Platform Development
Budget: £20 – £250 GBP
PROJECT TITLE:
AI Receptionist Web Platform (Voice, Chat, Admin AI, Billing, GCP)
PROJECT DESCRIPTION:
We are looking for an experienced full-stack developer or a small professional team to build a secure, cloud-based AI receptionist web platform.
The platform will be used by organisations such as:
- schools and colleges
- garages and automotive service centres
- dentists and clinics
- councils and public-sector organisations
- other service-based businesses
The AI receptionist will handle voice and text conversations, collect information, store and analyse data, and assist administrators through an AI-powered admin panel.
This is a serious SaaS project. It is NOT a basic chatbot or a simple HTML website.
---
SYSTEM OVERVIEW:
The system will:
- run on Google Cloud Platform (GCP)
- use Cloud Run for backend services
- integrate OpenAI APIs for AI reasoning and speech
- integrate Vonage Voice API for outbound calls
- support prepaid usage-based billing
- provide strong data isolation and privacy
Each client (organisation) must have fully separated data storage and configuration.
---
TECHNICAL REQUIREMENTS:
BACKEND:
- Google Cloud Platform (GCP)
- Cloud Run (container-based, stateless services)
- Backend language: Node.js or Python (FastAPI preferred)
- REST APIs and streaming endpoints for voice/chat
- Secure authentication for admin users
AI & SPEECH:
- OpenAI API:
- Chat / reasoning
- Speech-to-Text
- Text-to-Speech
- AI behaviour must follow client-defined instructions
- Admin AI analyst must query stored conversations and uploads
VOICE:
- Vonage Voice API
- Outbound calls to UK mobile numbers
- Used mainly for low-balance reminder calls
- Internet-based voice support (WebRTC acceptable)
DATABASE & STORAGE:
- Separate database per client (Cloud SQL or Firestore)
- Encrypted at rest
- Cloud Storage for file uploads
- Maximum upload size: 100MB
- AI-searchable data (vector search or equivalent approach)
FRONTEND:
- Mobile-first responsive web application
- React / Next.js / Vue (open to suggestion)
- Voice chat interface
- Text chat interface
- Live transcript displayed during voice calls
- File upload interface
- Admin dashboard UI
---
USER ROLES:
1) PUBLIC USER / CALLER:
- No login required
- Can speak via voice in browser
- Can type messages in chat
- Sees live transcript of conversation
- Can upload documents, images, or video
- AI receptionist asks relevant questions and collects data
2) CLIENT ADMIN:
- Secure login
- Admin dashboard
- Ability to define AI instructions:
(e.g. how the AI receptionist should behave, what questions to ask)
- View conversations and summaries
- Ask Admin AI analyst questions (up to 50 queries per day)
- View statistics and insights
- View usage and remaining balance
- Top up prepaid balance (£10–£500)
- Request refund of unused balance (minus banking/transfer fees)
---
ADMIN AI ANALYST:
- Text-based AI assistant available only in admin panel
- Searches across:
- conversation transcripts
- uploaded files
- AI-generated summaries
- Example questions:
- “How many students reported sick today?”
- “List all bookings made this week”
- Returns summaries, counts, and insights
- Limited to 50 queries per day per client
---
BILLING & BALANCE LOGIC:
- Prepaid wallet system per client
- Balance checked before every AI request
- Usage deducted automatically
- If balance drops below 10%:
- system triggers automatic voice call to client
- If balance reaches £0:
- service is suspended
- Service resumes automatically after top-up
- Unused balance refundable on request (minus bank transfer fees)
---
LOW-BALANCE REMINDER CALL:
- Automatic outbound call to client’s phone number
- Uses Vonage Voice API
- Short AI voice message, for example:
“Your AI receptionist balance is running low. Please top up to avoid service interruption.”
- Must not spam:
- maximum once per 24 hours
- Log only:
- timestamp
- success or failure status
- No conversation content stored for these calls
---
SECURITY & PRIVACY REQUIREMENTS:
- Each client’s data must be isolated from all others
- All data encrypted in transit and at rest
- Platform owner must not view client conversations or uploads
- Minimal logging (no message or file content in logs)
- Role-based access control
- GDPR-aware design (suitable for schools and councils)
---
LANGUAGE SUPPORT:
- System must support multiple languages
- English required
- Must handle accents and imperfect speech
- Ability to switch languages per client or per conversation
---
DELIVERABLES:
- Fully working web application
- Deployed on Google Cloud Platform
- Clean, readable, and documented code
- Infrastructure and deployment instructions
- Basic admin usage guide
- Git repository access
- Support for post-delivery bug fixes (initial period)
---
NICE TO HAVE (OPTIONAL):
- WebRTC experience
- Vector databases (pgvector, similar)
- Terraform or Infrastructure as Code
- SaaS billing experience
- Experience with GDPR or public-sector projects
---
WHO SHOULD APPLY:
- Senior full-stack developers
- Cloud-native engineers
- Developers with real API, AI, and backend experience
PLEASE DO NOT APPLY if:
- you only build static websites
- you have no cloud or backend experience
- you rely purely on templates
---
PROPOSAL REQUIREMENTS:
Please include:
1) Relevant past projects
2) Your role in those projects
3) High-level architecture proposal
4) Estimated timeline (phases)
5) Estimated cost (fixed or milestone-based)
6) Preferred tech stack
7) Any risks or questions you see
---
TIMELINE:
- MVP target: 8–12 weeks
- Possibility of long-term collaboration after MVP
AI Receptionist Web Platform (Voice, Chat, Admin AI, Billing, GCP)
PROJECT DESCRIPTION:
We are looking for an experienced full-stack developer or a small professional team to build a secure, cloud-based AI receptionist web platform.
The platform will be used by organisations such as:
- schools and colleges
- garages and automotive service centres
- dentists and clinics
- councils and public-sector organisations
- other service-based businesses
The AI receptionist will handle voice and text conversations, collect information, store and analyse data, and assist administrators through an AI-powered admin panel.
This is a serious SaaS project. It is NOT a basic chatbot or a simple HTML website.
---
SYSTEM OVERVIEW:
The system will:
- run on Google Cloud Platform (GCP)
- use Cloud Run for backend services
- integrate OpenAI APIs for AI reasoning and speech
- integrate Vonage Voice API for outbound calls
- support prepaid usage-based billing
- provide strong data isolation and privacy
Each client (organisation) must have fully separated data storage and configuration.
---
TECHNICAL REQUIREMENTS:
BACKEND:
- Google Cloud Platform (GCP)
- Cloud Run (container-based, stateless services)
- Backend language: Node.js or Python (FastAPI preferred)
- REST APIs and streaming endpoints for voice/chat
- Secure authentication for admin users
AI & SPEECH:
- OpenAI API:
- Chat / reasoning
- Speech-to-Text
- Text-to-Speech
- AI behaviour must follow client-defined instructions
- Admin AI analyst must query stored conversations and uploads
VOICE:
- Vonage Voice API
- Outbound calls to UK mobile numbers
- Used mainly for low-balance reminder calls
- Internet-based voice support (WebRTC acceptable)
DATABASE & STORAGE:
- Separate database per client (Cloud SQL or Firestore)
- Encrypted at rest
- Cloud Storage for file uploads
- Maximum upload size: 100MB
- AI-searchable data (vector search or equivalent approach)
FRONTEND:
- Mobile-first responsive web application
- React / Next.js / Vue (open to suggestion)
- Voice chat interface
- Text chat interface
- Live transcript displayed during voice calls
- File upload interface
- Admin dashboard UI
---
USER ROLES:
1) PUBLIC USER / CALLER:
- No login required
- Can speak via voice in browser
- Can type messages in chat
- Sees live transcript of conversation
- Can upload documents, images, or video
- AI receptionist asks relevant questions and collects data
2) CLIENT ADMIN:
- Secure login
- Admin dashboard
- Ability to define AI instructions:
(e.g. how the AI receptionist should behave, what questions to ask)
- View conversations and summaries
- Ask Admin AI analyst questions (up to 50 queries per day)
- View statistics and insights
- View usage and remaining balance
- Top up prepaid balance (£10–£500)
- Request refund of unused balance (minus banking/transfer fees)
---
ADMIN AI ANALYST:
- Text-based AI assistant available only in admin panel
- Searches across:
- conversation transcripts
- uploaded files
- AI-generated summaries
- Example questions:
- “How many students reported sick today?”
- “List all bookings made this week”
- Returns summaries, counts, and insights
- Limited to 50 queries per day per client
---
BILLING & BALANCE LOGIC:
- Prepaid wallet system per client
- Balance checked before every AI request
- Usage deducted automatically
- If balance drops below 10%:
- system triggers automatic voice call to client
- If balance reaches £0:
- service is suspended
- Service resumes automatically after top-up
- Unused balance refundable on request (minus bank transfer fees)
---
LOW-BALANCE REMINDER CALL:
- Automatic outbound call to client’s phone number
- Uses Vonage Voice API
- Short AI voice message, for example:
“Your AI receptionist balance is running low. Please top up to avoid service interruption.”
- Must not spam:
- maximum once per 24 hours
- Log only:
- timestamp
- success or failure status
- No conversation content stored for these calls
---
SECURITY & PRIVACY REQUIREMENTS:
- Each client’s data must be isolated from all others
- All data encrypted in transit and at rest
- Platform owner must not view client conversations or uploads
- Minimal logging (no message or file content in logs)
- Role-based access control
- GDPR-aware design (suitable for schools and councils)
---
LANGUAGE SUPPORT:
- System must support multiple languages
- English required
- Must handle accents and imperfect speech
- Ability to switch languages per client or per conversation
---
DELIVERABLES:
- Fully working web application
- Deployed on Google Cloud Platform
- Clean, readable, and documented code
- Infrastructure and deployment instructions
- Basic admin usage guide
- Git repository access
- Support for post-delivery bug fixes (initial period)
---
NICE TO HAVE (OPTIONAL):
- WebRTC experience
- Vector databases (pgvector, similar)
- Terraform or Infrastructure as Code
- SaaS billing experience
- Experience with GDPR or public-sector projects
---
WHO SHOULD APPLY:
- Senior full-stack developers
- Cloud-native engineers
- Developers with real API, AI, and backend experience
PLEASE DO NOT APPLY if:
- you only build static websites
- you have no cloud or backend experience
- you rely purely on templates
---
PROPOSAL REQUIREMENTS:
Please include:
1) Relevant past projects
2) Your role in those projects
3) High-level architecture proposal
4) Estimated timeline (phases)
5) Estimated cost (fixed or milestone-based)
6) Preferred tech stack
7) Any risks or questions you see
---
TIMELINE:
- MVP target: 8–12 weeks
- Possibility of long-term collaboration after MVP