David Arnoux

can't recommend Agent 4 enough... test it for free with 20usd of credits here

replit giving us promethean fire again... I've had early access to this and it's pretty astounding... Basically you can now plan, design and build all at the same time so you're not waiting on agents to finish building one part before running the next. You're running multiple chat threads to plan and build different parts of your app in parallel..... All within Replit's amazing environment. Do do do test this out.

I'm releasing this new keyboard soon for @claudeai #claudecode power users. DM me for pre-orders.

@levelsio wait levelsio isn’t on claude code??

I really need Claude app to add image generation, because I used it a lot with ChatGPT to make quick images

MacOS in 2027

calling it

@aryanlabde viralbrain.ai

Why is nobody building tools that help with marketing? Since that is what most founders here would pay for.

@GeminiApp donkey shed I’m building. Needed all steps and materials. All there in one shot

What are you making with Nano Banana Pro? Drop your creations in the replies ↓

gemini 3 is wild. I’m currently building a new donkey shed. This is one shot…

the timing doesn’t get much better than this. build this now. In 2024–25, lots of companies rushed to build smart AI “helpers” using tools like LangChain and AutoGPT. In tests and demos, everything looked perfect. But in real life, these helpers keep messing up. Weird real-world situations confuse them, some tools they use don’t always work, and important tasks get stuck: insurance bots stop finishing claims, shopping bots freeze at checkout, and support bots get stuck in endless loops. build a post-launch rescue layer for agentic AI: a lightweight SaaS that monitors full agent workflows end-to-end, auto-generates clear failing examples when things break, explains issues in plain language with step-by-step fixes for non-technical owners, plugs into LangChain / AutoGPT with fast connectors, and can optionally auto-fix repeatable failures like timeouts, missing fields, or known bad tools. first to ship workflow-first monitoring + edge-case generation + non-technical triage becomes the default safety layer for agentic AI in finance/insurance/healthcare/ecom. timing is perfect. reply “RESCUE” and i’ll dm a full breakdown: mvp scope, integrations, pricing, and go-to-market.

an opportunity for small restaurants and local chains. build this now. restaurant owners are busy. the daily special doesn’t get posted online, so fewer people show up. they need a simple, low-cost app that writes chef-style posts and auto-schedules them to instagram and facebook. build a daily-specials poster for restaurants that makes a daily calendar. write local posts from the menu and events. one-tap approve or edit. auto-schedule. pick safe, on-brand photos. show simple results like calls, reservations, and covers. why now: better language models. stable scheduling apis. big, growing social market with many restaurants still unserved. pricing and metrics: $15/month. ltv $700. ltv:cac 3.5:1. payback 5 months. ttm 8 months. timing is right. first to ship a daily-specials poster for restaurants wins local markets. drop a comment and i will dm you the full report with execution plan, mvp scope, partner targets, revenue model, and metrics.

an opportunity for small restaurants and local chains. build this now. restaurant owners are busy. the daily special doesn’t get posted online, so fewer people show up. they need a simple, low-cost app that writes chef-style posts and auto-schedules them to instagram and facebook. build a daily-specials poster for restaurants that makes a daily calendar. write local posts from the menu and events. one-tap approve or edit. auto-schedule. pick safe, on-brand photos. show simple results like calls, reservations, and covers. why now: better language models. stable scheduling apis. big, growing social market with many restaurants still unserved. pricing and metrics: $15/month. ltv $700. ltv:cac 3.5:1. payback 5 months. ttm 8 months. timing is right. first to ship a daily-specials poster for restaurants wins local markets. drop a comment and i will dm you the full report with execution plan, mvp scope, partner targets, revenue model, and metrics.

an opportunity for general contractors and project teams. build this now. rfis (project questions) are slow and messy. project managers spend time hunting emails and files instead of making decisions. the need: a smart helper that reads the question, pulls the facts, writes a quick draft answer, and sends it to the right person. build it: pull new questions from email and project apps. pick out key facts like plan page, location, trade, dates, and blockers. suggest a short reply using past answers and specs. send to the right role with a clear deadline and gentle reminders. keep everything in a simple search. save a history of who did what and who can see it. why now: better ocr and language models. stable cloud connectors. teams already in the cloud want faster decisions. pricing and metrics: ltv $6,000. ltv to cac 4 to 1. payback 9 months. time to market 12 months. timing is right. first to ship a focused rfi resolver wins pilots. drop a comment and i will dm the full report with execution plan, mvp scope, integration map, revenue model, and metrics.

someone should build this. the timing is perfect. sales reps, support agents, consultants, and founders paste customer info into public ai chats to get quick help. that can leak private data. there is a need for a tiny browser add-on that checks text before send, hides or blocks private bits, runs on your computer only, and follows simple team rules. build a browser extension that scans text fields and the clipboard. detect emails, phone numbers, ids, and account numbers. redact, warn, or block before submit. run fully on the device so nothing is uploaded. add simple admin rules and light reporting. today’s tools cover the hard parts. solid pii detectors. extension apis. local storage and policy controls. ship: install in minutes. block risky pastes on popular chat sites. one click override with a reason. team dashboard with anonymized alerts. pricing: team plans with usage. ltv 360 dollars. ltv to cac 4 to 1. payback 5 months. time to market 3 months. timing is right. first to ship a simple, local browser guard wins fast adoption. drop a comment and i will dm the full report with execution plan, mvp scope, buyer targets, revenue model, and metrics.

doing tax paperwork by hand is slow. people tag each invoice and send it to the right person, and files get lost. problems are missed. we need a simple online tool that reads e-invoices as they arrive through the Peppol pipe, adds the right tax labels, keeps a live “audit folder” with all proof, and warns you early if something looks wrong. build it: direct peppol + e-invoice ingest, auto vat mapping, live dossier with evidence links, risk scoring and exceptions, dashboards for reconciliation and deadlines, fast erp integrations, central rule updates. why now: better ocr and faster language models, peppol + vida deadlines, budgets moving to compliance tools. pricing and metrics: tiers per entity and volume. ltv $200k, ltv:cac 8:1, payback 9 months, ttm 32 months. timing is right. first to ship near real time vat reporting with audit-ready dossiers wins pilots and renewals. drop a comment and i’ll dm you the pilot plan, mvp scope, partner targets, pricing model, and metrics.

an opportunity for SMBs and creator tools. build this now. small businesses want to use fans’ photos and short videos as ads. But it’s slow and risky. Clips show up in DMs. someone has to check each one by hand to make sure it’s okay. then they must make different versions for TikTok, Instagram, and YouTube, which takes hours. they miss the right moment and spend too much. why now: cheaper multimodal gpt-4 apis. bigger creator spend. app store plugins for Shopify and Intercom. brands want safe, authentic ads at low cost. build a fan-made ad factory: fans upload. brands review. the system cleans rights, auto-formats for TikTok, Reels, YouTube, and Stories. add captions, brand colors, and safe-language checks. one tap to publish and track performance. today’s tools cover the hard parts. image and video formatting. brand-safety filters. link-in plugins. simple payouts and rights screens. ship: upload to review to publish in minutes. templates by channel. clear consent receipts. live metrics per ad. pilot with local stores and events. show faster ad creation and lower cost per result. pricing: $15 per month with premium analytics. ltv 270 dollars. ltv to cac 3 to 1. payback 9 months. time to market 6 months. timing is perfect. first to ship brand-safe fan ads for SMBs wins local markets. drop a comment and i will dm the full report with execution plan, mvp scope, partner targets, revenue model, and metrics.

@pmitu @Replit

Hey, vibe coders, what's your choice: - GitHub Copilot - Claude Code - ChatGPT - Windsurf - Lovable - Cursor - Bolt - v0 - ? 👨‍💻 Choose your fighter

been testing this for a few days. my PAL went from helpful tool → someone who knows my context → actually anticipating what i need next. first time i've caught myself saying "talk to you tomorrow" to an AI. that's new.

an opportunity for corporate development. build this now. files get mis-tagged and sent to the wrong reviewer. deals slow and mistakes slip through. there is a need for a document router that reads uploads, adds clear tags, and routes to the right person with an audit trail. why now: better ocr and faster language models. stable doc and dealroom apis. pressure to close faster with lean teams. build a document router. read PDFs and scans. tag “invoice,” “contract,” “cap table” with confidence. route by role, phase, and deadline. queue low-confidence files. track time saved and errors avoided. plug into dealrooms, DMS, email, and slack. today’s tools cover the hard parts: robust ocr, retrieval, rules engines, SSO, and webhooks. ship: auto-tag on upload. one-click route. exceptions queue. search by party, date, and amount. export logs for counsel and finance. pricing: per dealroom or per user. LTV $3,000. LTV to CAC 3.75 to 1. payback 8 months. time to market 12 months. timing is perfect. first to ship fast, accurate auto-tagging and routing wins pilots and renewals. drop a comment and i will dm the full report with execution plan, mvp scope, integration map, revenue model, and metrics.

an opportunity for e-commerce and saas teams. build this now. agents waste hours rewriting replies. bots answer in one flat voice and customers bounce. there is a need for a no code voice router that makes the bot speak in the right tone for each customer, page, and issue. llms are better at style. app stores and plugins make installs fast. leaders want consistent, human replies across channels. build a brand voice chat router. set voices once. route by segment, channel, page, and topic. pick the best model per voice. show short receipts so teams see which rule fired. today’s tools cover the hard parts. stable chat apis. marketplace plugins. guardrails for refunds and pii. usage analytics. ship: connect intercom or shopify, import macros, create 3 voices, add 5 rules, a/b on checkout and billing, show csat lift and fewer escalations. pricing: team plans with usage. ltv 15,000 dollars. ltv to cac 4 to 1. payback 11 months. time to market 8 months. timing is perfect. first to ship simple, reliable voice routing wins pilots and renewals. drop a comment and i will dm the full report with execution plan, mvp scope, integration map, revenue model, and metrics.