Ideas Lab

Real problems people complain about online, pulled every morning and scored out of 100. Build, validate, or skip. How scoring works

Ideas
4,677
BUILD
699
Sources
8
Last sync
4h ago
4,677 ideas

How much should I pay for the given website.

Project Title: Micro-Task Platform Project Overview Build a web platform where clients can create online tasks and workers can complete them to earn money. Workers submit proof, tasks are verified automatically or manually, and earnings are credited to their wallet. The platform must support crypto withdrawals and an admin management system. User Roles 1. Worker - Create account and login - View available tasks - Complete tasks and submit proof (link or screenshot) - Track earnings and task history - Request withdrawals 2. Client - Create campaigns/tasks - Set reward amount per task - Define number of workers required - Track task completion statistics - Pay for campaigns 3. Admin - Approve or reject task submissions - Manage users - Manage campaigns - Process withdrawals - Monitor platform activity Core Features User System - Registration and login - Email verification - User dashboard - Wallet balance display Task System - Task listing page - Task instructions page - Proof submission (URL or screenshot upload) - Task status tracking (pending / approved / rejected) Verification System - Automatic link validation for tasks when possible - Manual admin approval for complex tasks Wallet System - User balance management - Earnings credited after task approval - Withdrawal request system Payments - Crypto payments (USDT, BTC, ETH) - Admin manual approval for withdrawals Admin Dashboard - View all users - Manage tasks - Approve submissions - Approve withdrawals - Platform analytics Security & Anti-Fraud - Duplicate submission detection - Rate limiting for tasks - IP monitoring for multiple accounts Technical Requirements Frontend Responsive web interface. Backend API based architecture. Database Store users, tasks, submissions, and transactions. Optional Features (Phase 2) - Automated verification via social platform APIs - Referral system - Reputation score for workers - Automated crypto payouts - Real-time notifications

Reddit7mo agoToolDev

59FL score
1Sign in to vote

outsourced our phone answering to save time and now we're spending more time fixing the notes than we saved

We run a small insurance agency and phones are the biggest time suck in the office. Figured we'd try one of those call answering services to handle the easy stuff while we focus on quotes and renewals. Three weeks in it's a mess. The operators keep mixing up policy numbers and claim numbers in the notes they send us, carrier names are misspelled half the time, and if a client asks anything slightly specific they just transfer the call right back to us. So we're paying for a service that handles maybe 20% of calls on its own and creates cleanup work for the rest. One call really got me. Commercial client called about adding a location to their bop. Notes came back saying "customer wants to add a body shop to their policy." I still don't understand how you get there from what was actually said. Another time someone called about their personal umbrella policy and the only note was "customer has question about umbrella insurance." Zero context, zero details, so we had to call the client back and redo the whole conversation. The frustrating part is the terminology isn't even that complicated once you learn it but these generic services clearly don't train their people on any of it. I'm pulling the plug this week. Been poking around at some of the industry specific options, saw a few names come up like sonant and a couple others that are supposedly trained on insurance calls specifically but honestly at this point I'm skeptical anything that isn't our own staff will get it right. Anyone else dealt with this when trying to automate or outsource something in a specialized field? feels like the generic tools just don't cut it when there's real domain knowledge involved.

Reddit7mo agoToolProductivity

78FL score
1Sign in to vote

Respan (YC W24) – Open-source Rust AI gateway built around plugins

Hey HN, this is Andy from Respan (https://www.respan.ai). We are building an open-source Rust AI gateway. It is a decoupled gateway, so everything is a plugin, including different providers, protocols (MCP, voice), modalities (images, videos), and gateway features (fallbacks, PII redaction, smart routing).Instead of packaging tons of providers and features into one gateway, Respan provides a small core and public interfaces that teams can use to assemble the AI infrastructure their application needs. That way, users can build their own features as additional plugins without having to change the gateway's core.Here’s a video demo: https://www.youtube.com/watch?v=pVv2ibXiTXMThere are countless LLM gateways and most of them solve a common problem: calling different providers from a single endpoint. However, adding new features can become prohibitively difficult. So developers can end up building another gateway to solve a specific problem.We've been building our gateway based on LiteLLM since 2024, and now it's providing services for thousands of orgs. Each org needs different API key settings, routes, and limits. To support that, we kept wrapping LiteLLM, adding more features to it, and at one stage, our wrapper code exceeded 100,000 lines.This January, a routine LiteLLM upgrade caused intermittent connection errors, and we had to roll back. We still haven't figured out the cause, but since then, we no longer upgrade LiteLLM directly and instead manually backport upstream fixes. These experiences made us realize that continuing to patch LiteLLM is not a sustainable long-term solution, and we had to redesign the gateway ourselves.While redesigning our gateway, we wanted to make it easier to add features without breaking existing ones. Inspired by Nginx, we think an AI gateway should have a really small core, only responsible for dispatching events.We believe that everything should be a plugin in our gateway, including provid

Hacker News19h agoToolAI

78FL score
0Sign in to vote