We are seeking a skilled developer to create a SaaS web application that includes an AI-powered blog generator integrated with n8n. The ideal candidate will have experience in both SaaS development and AI technologies. Your role will involve designing the architecture, implementing the blog generation features, and ensuring seamless integration with n8n for automation workflows. Strong problem-solving skills and the ability to work collaboratively are essential for this project.

I want a skilled developer/team to build a modern web application (SaaS) that enables my clients to easily generate and publish AI-written, SEO-optimized blog posts—powered by n8n automations in the backend.
Scope:
• User login & dashboard
• Simple web form for inputting blog topics/keywords
• Integration with n8n workflow via secure webhook or API
• n8n handles AI content generation (OpenAI/GPT-4), AI image creation (e.g. DALL-E), and automatic posting to WordPress
• Status & results feedback in dashboard
• Stripe (or similar) payment integration for subscriptions
• Admin panel to manage users/plans

SaaS Subscription Model
Pricing Tiers:
• Starter: €29/maand - 10 blog posts
• Professional: €79/maand - 50 blog posts
• Enterprise: €199/maand - Unlimited posts + white-label

Relevant skills:
- SaaS Development
- AI / Machine Learning
- n8n Integration
- Web Application Development
- JavaScript / TypeScript
- RESTful APIs

We’re building an AI phone receptionist for hair salons — an MVP that actually works and can handle the first 500 real users.

The Product (MVP Scope)
A phone-based AI receptionist that:

Lets salon owners sign up via their smartphone

Allows them to:

Take a photo of their pricing

Set opening hours

Input basic salon info (e.g., services, phone number)

Then, instantly handles:

Appointment booking, canceling, rebooking

SMS reminders

voice call handoff or voicemail capture

This is not a demo — we want a real, working MVP for actual salon owners.

What We’re Looking For
You build fast using AI tools (Cursor, Copilot, GPT, Claude, etc.)

You know how to ship production-ready MVPs that don’t break

You can suggest a sensible stack and keep infra simple (e.g. Firebase, Supabase, Railway)

Bonus if you’ve handled telephony (Twilio / Plivo) or calendar scheduling

We care about clean code, speed, and being able to scale later.

Deliverables
Functional web/mobile-friendly MVP
Auth + onboarding flow
Admin panel (basic) for viewing/editing bookings
Simple deployment (Vercel, or similar)
implementation of easy call forwarding solution (open to suggestions)
This is an exciting build with real use cases and potential for long-term work.

Apply with:

1–2 examples of MVPs you built fast

Whether you use AI tools (and how)

Your proposed stack for this build

We are building a web-based platform that includes a dynamic media player, user dashboards, and admin panel.

This is a standalone project with potential for long-term work. We need a reliable full-stack developer to help bring the MVP to life.

I would love if we can speak spanish

We are building an institutional defi platform for the issuance, trading, and management of security tokens and DeFi products, built on the XRPL (XRP Ledger).

Our current MVP is live (backend in .NET, frontend in Razor Pages), with key components like KYC, payments and basic database already functioning. We are now transitioning to a React-based full-stack production system.

We are looking for an experienced development partner or team to execute the full development roadmap over 3 months (6 milestones, each 2 weeks).

Tech Stack - Must-have:
+ React 18 (Vite, TypeScript, Tailwind)
+ XRPL / xrpl.js (XLS-20, AMM, Hooks)
+ PostgreSQL
+ Node.js or .NET backend integration
+ REST / GraphQL API handling
+ Web3Auth integration
+ Firebase / Push Notifications
+ Docker & CI (GitHub Actions)
+ Stripe, NowPaymentsn Transakt (On/Off ramp)

Nice to have:
+ Knowledge of XLS-20 token standards & metadata
+ DeFi UX experience (DEX, AMM, staking flows)
+ XRPL Hooks (written in C)
+ Experience working with fintech or institutional apps

Scope of Work (Milestones Overview)
Each milestone will be implemented in 2-week sprints:

Milestone 1: Wallet + Token Distribution
+ Integrate Web3Auth (email + social)
+ Build checkout-to-wallet token distribution flow (XRPL XLS-20)
+ Create wallet dashboard with KYC gating and transaction records

Milestone 2: Mobile PWA & Unified Investment Flow
+ Setup mobile PWA shell (React-based)
+ Finalize Web + Mobile investment logic, KYC verification, and payment UX
+ Deploy push notifications via Firebase
+ Final deployment, bug fixes

Milestone 3: (T99) Utility Token, Tiered Access & Staking
+ Create (T99) utility token swap interface and staking engine
+ Implement logic for issuer/investor tier upgrades
+ Enforce gated access to platform features via staking
+ Auto reward distribution + admin panel overrides

Milestone 4: Security Token Launch Engine (for Issuers)
+ XLS-20 minting wizard for different token types (Equity, ART, RE, Bond, Index)
+ Document uploads, NAV input, dividend hooks, redemption rights
+ Billing engine for monthly fees (based on token type and tier)
+ Issuer dashboard with cap table, trustlines, lifecycle control

Milestone 5: DEX, AMM & Redemption
+ Implement orderbook DEX screen with chart, swap, trustline prompt
+ AMM liquidity pool interface (add/remove LP)
+ Portfolio analytics, token stats, APRs, redemption + dividend history

Milestone 6: Admin Panel + Compliance Management
+ Admin dashboard to manage asset classifications, billing lock logic
+ Asset-specific compliance rules
+ Security monitoring & override flows

Deliverables
1. Fully functional React + XRPL PWA
2. XRPL trustline automation and XLS-20 token distribution
3. Responsive mobile-first layout
4. Admin panel and issuer dashboards
5. Webhooks for payment → mint → distribution flow
6. Live deployment (staging + production)
7. QA and documentation (basic) per milestone

What We Look For:
+ You/your agency must be highly experienced in blockchain-based dApp development
+ Proven track record of XRPL or similar L1 DeFi systems
+ Strong full-stack experience with React + Node/.NET + PostgreSQL
+ Fluent English & ability to collaborate asynchronously
+ High responsiveness, ability to hit 2-week sprint targets
+ Clean, well-commented code and test coverage

Timeline: 3 months total (6 milestones x 2 weeks)
Start Date: August 2025
Budget: Hourly ($25–40/hour) or fixed per milestone (est. €15K–€25K total budget)

To apply, please share:
- Relevant XRPL / React work (with links)
- Your team structure (if agency)
- Your availability per week
- Any questions or clarifications you’d like before starting

We are looking for a long-term development partner, as the platform will require ongoing feature expansion and optimization. New milestones are already planned for Q1–Q2 2026, including advanced issuer tools, institutional APIs, and XRPL-based settlement enhancements.

Hello builder! 👋
I'm on the hunt for a skilled developer to help bring to life a new AI-driven content tool. The goal? To create a powerful web app that generates high-performing long YouTube scripts (3000+ words), tailored for audience retention, storytelling youtube principles and specific persona.

🔧 What We’re Building
Imagine a tool that writes YouTube scripts, but not just any scripts. It learns from real performance data (like audience retention, drop-off points, watch time, etc.) and uses that feedback loop to improve future scripts automatically.

Check out this related SaaS concept for inspiration:
https://arvow.com/ or https://writesonic.com/

🧱 Tech Stack
We’re looking for someone with proven experience using:
Next.js (React-based frontend)
Vercel (for deployment and hosting)
Supabase (as the backend/database layer)

Bonus if you have experience integrating with:
YouTube APIs (Analytics, Reporting)
Machine learning workflows (optional for now)

🧠 Ideal Candidate
You’ve built (or contributed to) a text generation or content-writing app
You’re familiar with AI/LLMs (we’ll handle most of the ML logic, but your experience helps)
You write clean, modular, scalable code
You can collaborate async, communicate clearly, and bring creative suggestions

📣 Interested?
If you’ve built something similar, send me a message with examples or a portfolio.
Know someone who fits this? Feel free to forward this post!

*MUST: Please, specify the link apps you have previously built or worked on. And specify your role on those. If this information is not attached the proposal won't be read.

Let’s build the future Youtube App Mr Beast will use, trust!

We are seeking a no/low-code freelancer to assist in developing and automating a custom booking system using Airtable, Softr, n8n, and Make for guided excursions. The ideal candidate will have experience in these platforms and a passion for creating seamless user experiences. Your role will involve setting up and optimizing workflows, integrating the necessary tools, and ensuring everything operates smoothly for our clients. If you're skilled in building user-friendly systems with minimal coding, we want to hear from you!

Dear freelancers,
I'm looking for a talented full stack developer with expertise in artificial intelligence to help me build a fully functional autonomous agent using available open source projects.

The work will include everything from server setup to finished product with front and backend and integrated payment/subscription feature.

Please propose a solid working plan. I intend to use Falcon 40B as LLM model but I'm open to suggestions.

To confirm that you have carefully read the job description please start your proposal with the word UMBRELLA.

Best regards
Jacques

Hello builder! 👋
I'm on the hunt for a skilled developer to help bring to life a new AI-driven content tool. The goal? To create a powerful web app that generates high-performing long YouTube scripts (3000+ words), tailored for audience retention, storytelling youtube principles and specific persona.

🔧 What We’re Building
Imagine a tool that writes YouTube scripts, but not just any scripts. It learns from real performance data (like audience retention, drop-off points, watch time, etc.) and uses that feedback loop to improve future scripts automatically.

Check out this related SaaS concept for inspiration:
https://arvow.com/ or https://writesonic.com/

🧱 Tech Stack
We’re looking for someone with proven experience using:
Next.js (React-based frontend)
Vercel (for deployment and hosting)
Supabase (as the backend/database layer)

Bonus if you have experience integrating with:
YouTube APIs (Analytics, Reporting)
Machine learning workflows (optional for now)

🧠 Ideal Candidate
You’ve built (or contributed to) a text generation or content-writing app
You’re familiar with AI/LLMs (we’ll handle most of the ML logic, but your experience helps)
You write clean, modular, scalable code
You can collaborate async, communicate clearly, and bring creative suggestions

📣 Interested?
If you’ve built something similar, send me a message with examples or a portfolio.
Know someone who fits this? Feel free to forward this post!

Let’s build the future Youtube App Mr Beast will use, trust!

Project Overview: AI-Powered Preventive Health Coach Embedded in a Moodle-Based Learning Ecosystem

1. Vision & Purpose
You are being invited to help build a next-generation preventive health education platform that combines science-backed lifestyle medicine, emotionally intelligent coaching, and personalized AI guidance in a scalable, compliant, future-proof architecture. The goal is to support individuals on health journeys—improving metabolic health, sleep, stress resilience, mindset, longevity, and daily habits—by marrying structured curriculum with a real-time AI companion that understands their progress, context, and needs.
This platform is not just another course delivery site. It is a hybrid learning + coaching system where evidence-based knowledge (from preventive medicine, nutritional science, exercise physiology, sleep science, and behavioral psychology) meets a kind, supportive AI coach that acts like a smart companion—reminding, nudging, adapting, and celebrating progress—while safely handling boundaries (never diagnosing, always escalating appropriately when needed).
The learner journey starts inside a Moodle course, where structured content (modules, exercises, lifestyle resets) is delivered, tracked, and certified. Layered on top is the AI Health Coach: a dynamic, personalized system that interprets the user’s learning status, self-reported health data, and historical interaction to deliver context-aware, motivational, actionable guidance. Over 2–3 years the architecture must allow the coach to be reused or extended into a standalone B2C mobile/web application, retaining all intelligence, history, and personalization.

2. High-Level Objectives
Primary Outcomes - Deliver personalized, adaptive coaching that helps users translate course material into sustainable behavior change. - Ensure that every recommendation is grounded in scientific evidence, tied back to course content when applicable, and delivered with empathy and clarity. - Establish trustworthy boundaries: the coach is supportive but not therapeutic; it defers diagnosis and escalation to qualified professionals when red flags appear. - Build a persistent memory per user so the system becomes more intelligent and tailored over time, recognizing past behaviors, preferences, and what interventions worked. - Integrate seamlessly with Moodle to leverage course state, deep-link into relevant lessons, and surface on-platform interventions. - Architect the solution for data protection
compliance in the Swiss/DACH context (nFADP/GDPR-style), particularly given the sensitivity of health-related data. - Design for model autonomy and evolution, such that dependence on external hosted models can be reduced over time and new frontends (like a custom app) can reuse the same intelligence.

3. Strategic Value
This system sits at the intersection of:
- Education: Structured, evidence-based lifelong learning delivered via Moodle.
- Behavioral Science: Psychology of change embedded in conversation and micro-action planning.
- Health Optimization: Lifestyle medicine for chronic prevention (e.g., metabolic reset, stress management, sleep hygiene, etc.).
- Technology & AI: A smart, adaptive coach that personalizes, remembers, and improves.
- Compliance & Trust: Data-handling standards that make it acceptable for partnerships with insurers and institutions in regulated markets.
That combination creates a defensible product: learners get real results, the platform can earn institutional credibility through tracking & certification, and the AI coach becomes a sticky, intelligent layer that differentiates and scales.

4. System Overview (Conceptual)
At a high level the system is composed of:

1. Moodle Learning Environment – Hosts the course curriculum: modules on sleep, nutrition, movement, stress, longevity; tracks completion, quizzes, competencies.

2. AI Health Coach Engine – Independent microservice that consumes user state and produces personalized advice and next actions.

3. Multi-source Knowledge Layer:
- ChatGPT (or equivalent) for delivery: persona, empathy, behavior-change framing, motivational language.
- MedGemma / Gemma self-hosted model for clinical reasoning and medical evidence synthesis.
- Primary literature retrieval (PubMed, leading institutions) and internal course content via a retrieval-augmented generation pipeline to avoid hallucination.
4. Persistent Memory – Stores user interaction history, preferences, interventions tried, and outcomes to tailor future coaching.
5. Integration Layer – Brings together learning context (from Moodle), health metrics (self-reported or tracked), and memory to assemble a sanitized, relevant user profile for the coach.
6. Frontend Embedding – An in-line coach interface inside Moodle for real-time interaction, feedback capture, and redirection to course content.
7. Compliance & Security Framework – Consent capture, pseudonymization, encryption, access auditing, red-flag detection, and user data lifecycle management.

5. Core Functional Expectations (What the AI Coach Actually Does)
- Understand the learner’s context: what modules they’ve done, where they are stuck, their health inputs (e.g., sleep quality, mood, weight trends), and their personal goals.
- Respond with a persona that is kind, supportive, empathetic, non-judgmental—acknowledging effort before suggesting the next step.
- Ground advice in:
- The course content (linking to relevant modules).
- Scientific evidence (via MedGemma and literature retrieval).
- Coaching psychology (via ChatGPT-style framing for motivation and habit initiation).
- Enforce boundaries: If a user shares something outside the intended coaching scope (relationship stress, mental health crises), the system shows empathy, pivots to immediate coping strategies (mindfulness, breathing), and if serious red flags appear, suggests seeking professional help.
- Personalize over time: Remember what advice worked, adapt phrasing to user preference, escalate or modulate intensity based on adherence and historical behavior.
- Collect feedback: Capture explicit (ratings) and implicit (compliance/outcomes) signals to refine future recommendations and build training data for evolution.
- Fail gracefully: If evidence is insufficient or a user query is ambiguous, provide transparent clarification (“I don’t have enough context to be sure—here’s a small experiment to try, and if it persists, here’s how to escalate.”).

6. Data Sources & Roles
- ChatGPT (or equivalent conversational LLM)
Role: Persona, empathy, conversational scaffolding, behavior change psychology, micro-planning, motivational reinforcement.
Constraint layer: Restrict topic drift; anchor responses to health program relevance; detect and manage out-of-scope disclosures.
- MedGemma / Gemma
Role: Clinical evidence synthesis, high-fidelity medical reasoning for lifestyle interventions (e.g., strategies to improve insulin sensitivity, evidence-based sleep hygiene), and flagging risk patterns.
Provides the scientific backbone that supports “why” a suggestion makes sense.
- Primary Literature (PubMed, Harvard, etc.)
Role: Supplement and update the knowledge base with the latest systematic reviews, meta-analyses, and domain-specific findings; indexed for retrieval to ground answers and increase credibility.
- Course Content
Role: Canonical source of interventions. Advice that maps to curriculum modules includes deep links and contextual operationalization (e.g., “See Module 5 for the full routine; try this 5-minute practice today.”).
- User Memory / Interaction History
Role: Enables adaptation, personalization, and continuity across sessions.

7. Compliance & Privacy Requirements (Strategic Imperative)
- Explicit consent management for all health-related personalization.
- Pseudonymization of identities; separate storage of PII and behavioral/health logs linked only by internal identifiers.
- Sanitization before external calls: Nothing that can identify the user (name, exact birthdate, raw sensitive medical records) is sent to third-party services unless absolutely necessary and with consent—only abstracted context goes out.
- Auditability: Every recommendation must be traceable—including which evidence/sources it referenced, what triggered a red-flag, and what feedback was provided.
- Data residency: Sensitive data lives in Switzerland or EU-hosted compliant infrastructure.
- User rights: Export, correction, and deletion must be supported for both the learning history and coaching memory.
- Escalation logic: Build-in red-flag detection with safe fallback language and referrals to qualified care when thresholds are met.

8. Future Evolution
- Engine swap capability: External dependencies (e.g., initial use of hosted ChatGPT) must be encapsulated behind an abstraction so they can be replaced by self-hosted or custom fine-tuned models in the future without changing the API or user-facing integration.
- Standalone App Reuse: After establishing proof-of-concept and traction inside Moodle, the same AI coach backend (context assembly, memory, personalization, evidence retrieval, safety logic) must be reusable in a completely new front-end (web/mobile app) to serve consumers directly.
- Learning loop: Data collected from interactions is not only used for immediate personalization but also forms the basis for gradual model improvement—either via refining prompt templates, training internal preference models, or fine-tuning open models.

9. Candidate Profile: Required Tools & Skills
This section describes the concrete capabilities the candidate must possess. Only candidates who can honestly say they have mastery or deep working experience in the core areas below should proceed; the detailed brief will be shared after vetting.

A. Core Technical Skills & Knowledge
1. Backend Development (API-first) - Strong experience building, deploying, and maintaining RESTful APIs with Python (FastAPI preferred) or Node.js (Express/Fastify). - Design of pluggable / swappable model integration layers (adapter patterns).

2. Large Language Models & Prompt Engineering - Hands-on experience integrating ChatGPT/OpenAI APIs or similar LLMs. - Skilled in prompt engineering for persona enforcement, behavior change framing, and safe output shaping. - Ability to implement scope filters to detect and manage out-of-scope input (e.g., using classifiers or heuristic parsing).

3. Open / Self-hosted Model Integration - Familiarity with models like MedGemma / Gemma or other domain-specialized open models; experience deploying and interfacing with them. - Understanding of building or integrating local inference stacks and fallbacks. - Knowledge of techniques for gradual autonomy: replacing external models with internal ones without breaking interface contracts.

4. Retrieval-Augmented Generation (RAG) - Experience with vector databases: Qdrant, Weaviate, Milvus, RedisVector, or equivalent. - Building pipelines to index heterogeneous content (course content, literature summaries, evidence snippets) and retrieve contextually relevant vectors to supplement generation.

5. Learning System Integration - Familiarity with xAPI and Learning Record Stores (LRS) (e.g., Learning Locker) or ability to consume structured learning data from Moodle via its APIs. - Understanding how to map course progress into upstream context for the coach.

6. Personalization & Memory Systems - Designing per-user memory layers: what to store, how to version, how to retrieve safely, how to use it to modulate future advice. - Experience avoiding leakage between user profiles (strict isolation).

7. Moodle Embedding / Integration - Experience embedding external widgets/APIs inside Moodle course pages. - Token/session propagation from Moodle to external systems securely. - Deep-link creation into lessons/modules.

8. Identity & Security - OAuth2 / JWT mechanisms: secure token handling and session management. - Experience with SSO flows or integrating with identity providers (e.g., Keycloak or equivalent).

9. Data Privacy & Compliance Engineering - Practical implementation of GDPR-style privacy controls: consent versioning, pseudonymization, deletions, export, audit logging. - Understanding of secure handling of sensitive health data (encryption-at-rest, in-transit, access controls). - Building red-flag detection and safe escalation flows.

10. DevOps & Infrastructure - Containerization (Docker) and ideally orchestration knowledge (Kubernetes or equivalent) for scaling components independently. - CI/CD pipelines, rollout/versioning strategies for model swaps. - Observability: logging, monitoring (performance, security), alerting. -

Experience with hosting in regulatory-compliant regions (Swiss/EU) and managing secrets (e.g., using Vault, KMS).

11. Database Design - Relational (PostgreSQL) or hybrid storage: schema separation for PII vs behavior data; encryption strategies. - Capability to scale read/write patterns for memory & context retrieval.

12. Evidence Provenance & Content Handling - Ability to ingest, structure, tag, and surface scientific evidence (guidelines, literature abstracts) in a way that’s queryable and can be attached to advice. - Familiarity with building or normalizing content using controlled vocabularies/ontologies (helpful but not strictly required at first).

B. Tooling & Ecosystem
- Git (version control, branching strategies).
- OpenAI API (or equivalent) integration.
- Vector DBs: Qdrant, Weaviate, or alternatives.
- Moodle platform (practical understanding of how to integrate outside services).
- Identity systems: JWT, OAuth2 flows, Keycloak or equivalent.
- Secure storage/encryption libraries.
- API documentation frameworks (OpenAPI/Swagger).
- Testing tools: unit/integration testing expertise, especially for data sanitization, persona output validation, safety logic.
- Observability stacks: ELK/Prometheus/Grafana or managed equivalents for logging/monitoring.
- Deployment automation: scripting blue/green or canary deployments for model updates.

C. Soft / Strategic Skills
- Strong sense of product thinking: able to translate health coaching objectives into technical implementation with user empathy.
- Clear communication: can explain trade-offs, design decisions, and potential risks, especially around safety and compliance.
- Documentation discipline: produce maintainable API specs, onboarding guides for future app integration, and model swap procedures.

- Data ethics mindset: conscious of bias, safety, user autonomy, and responsible AI practices.
- Collaborative: able to work with the client’s content/Moodle setup owner and evolve the coach jointly.

10. Nice-to-Have / Differentiators
- Prior experience with health coaching, lifestyle medicine, or preventive health tech.
- Familiarity with Swiss/EU compliance frameworks specifically (nFADP, HIPAA parallels, etc.).
- Experience fine-tuning LLMs via LoRA or parameter-efficient techniques.
- Exposure to human-in-the-loop systems for health-critical advice.
- Experience building adaptive dialogue systems that adjust based on user feedback and memory.
- Knowledge of semantic normalization (SNOMED CT, ICD, etc.) for better internal consistency (especially useful if scaling to more clinical depth).
11. Candidate Delivery Expectations (Screening Signals)
Before sharing the detailed brief, you should ask each candidate to demonstrate:
- A concrete example of a multi-source assistant they built or architected (LLM + external knowledge + personalization).
- How they handled persona enforcement and boundaries in a conversational AI (example prompts or system design).
- Their approach to data privacy in a health-related context (how they separate identity from behavior, consent capture, deletion).
- An outline of how they would design a memory system for per-user personalization without cross-leakage.
- A sketch of how they’d structure model abstraction to allow backend swaps (e.g., replacing ChatGPT with a self-hosted model).
- Experience embedding a service in a third-party platform (Moodle or similar) and propagating identity securely.

12. Summary
This project is the fusion of credible health education, intelligent coaching, and adaptive AI in a gated, compliant environment. The ideal candidate is not just a coder but a systems thinker who can:
- Fuse conversation, evidence, and curriculum into coherent, safe advice
- Build memory-driven personalization that feels human
- Maintain privacy and compliance as a first-class citizen
- Architect for evolution and reuse beyond the initial platform
If you have those capabilities and can speak concretely to how you’d implement the core pieces, you’ll be invited to receive the full detailed technical brief and proposal request.

We’re building an MVP for a lightweight tool called Bluebill Insight Suite, designed for mid-sized companies. The platform should allow users to upload invoices and contracts (PDF/CSV) and automatically generate insights using GPT and simple automation.

Scope – MVP Features:
1. SaaS Cost & Tool Detection

• Users upload invoice files
• GPT extracts software tools, pricing, anomalies, duplicates
• System suggests more cost-effective alternatives
• Flags possible Shadow IT usage

2. Contract Renewal Monitoring

• Users upload contracts (PDF)
• OCR + GPT extract start/end dates, renewal clauses, deadlines
• Email reminders before renewals
• Table view of contracts with status and recommendations

3. Cashflow Optimization

• Europe: Detect early payment discounts from invoice terms
• USA: Identify monthly billing and suggest switching to annual if savings are possible
• AI-generated tips for cost efficiency

Dashboard Requirements:
• Live savings overview
• Sorting and prioritization (e.g., cancel, renegotiate, pay early)
• Filter and search functionality

Tech Stack (Suggested):
• GPT-4 / OpenAI API
• Retool, Flutterflow, or Softr for the front end
• PDF.co or similar for document parsing
• Make or Zapier for automation workflows
• SendGrid or other tool for email triggers

Deliverables:
• Working MVP with a clear, modern UI
• Parsing logic and prompt setup
• Smart suggestions and automatic classification
• Dashboard with live savings tracking
• Basic admin settings or logic for region-based logic (EU vs US)

Please include:
• Estimated time to deliver
• Budget range
• Any previous experience with similar GPT or dashboard tools