0901 | Launch Roundup: Revolte, BrandJet, Particle Radar, StackScope

Mia: Welcome back to Product Hunt Daily on Bri Radio. I'm Mia, and I'm here with Milo.

Milo: Hey, good to be here. And we've got a fresh batch of launches today — from AI software delivery to podcast search and even a tool to plan your outfit.

Mia: Revolte is rolling out Interactive Sessions, StackScope promises technographics for the new web, and there's a speech-practice app with a solo developer behind it.

Milo: We'll walk through all of it, starting with Revolte's Interactive Sessions.

Mia: Revolte's new launch puts Interactive Sessions at the center, a new mode in its AI software-delivery platform. Co-founder and CEO Raj frames Revolte as one workspace for both interactive and autonomous AI development, built around a problem he says existing tools create: teams are forced to pick sides, where Cursor is hands-on and Devin is hands-off, and engineering leaders end up running Cursor, Copilot, and Devin as three separate tools with three governance models, three cost centers, and three procurement conversations. Revolte's pitch is not that it outperforms any one of them, but that teams no longer need all three.

Milo: Interactive Sessions itself is a tabbed workspace, one session per task, where an engineer drives the full software lifecycle step by step through architecture, code, tests, staging, preview, and deploy, approving every step. According to the maker, it needs no Jira setup, no CI configuration, no long onboarding, and getting from sign-up to a useful session takes under a minute.

Mia: The separate Autopilot mode, from Revolte's earlier launch, takes a Jira ticket and runs it end-to-end through pull request and deploy, with engineers approving the meaningful steps. Both modes share one governance layer: plan approval before code, inline diffs before merge, cost caps before deploy, and an audit trail on every action, all running on the same agent stack and Platform-as-Code contracts.

Milo: The target is engineering teams shipping production software, especially leaders trying to adopt AI across teams with different risk appetites, on one platform with one set of rules. On its website, Revolte says its AI agents keep the human in the approval loop on the one platform.

Mia: BrandJet launched on Product Hunt as a go-to-market platform that, according to its launch description, turns live public buying signals into outbound sales pipeline. It listens across X, Reddit, LinkedIn, and the open web for people asking about problems the user's product solves, then enriches each relevant signal into a lead with contact information and context, and reaches out on the channels prospects actually answer — email, LinkedIn, WhatsApp, X, Instagram, and Telegram.

Milo: Replies land in one inbox with a CRM behind it, and contracts can be signed in the same platform through a signing tool the company says it built, named Plutus. Co-founder Marsad said the team spent the last year building it and described BrandJet as one of the most comprehensive go-to-market platforms on the market, with the makers saying they use BrandJet to sell BrandJet. The launch offer is thirty percent off for the next three months with code BRANDJETHUNT30.

Mia: The whole platform also ships as a model context protocol, or MCP, server with more than seventy tools across listening, enrichment, outreach, inbox, and CRM, so outside AI agents can drive it; the makers' own agent, Artemis, runs on that same layer. In the launch discussion, commenters said combining buying signals, outreach, and CRM follow-up could save context switching for small sales teams, and one liked that outreach is tied to brand and competitor sentiment rather than treated as separate tasks.

Milo: One commenter also flagged the co-founder's comparison of ten thousand purchased leads, of which nine thousand nine hundred and forty were stale Gmail contacts, as evidence of how little value purchased lists tend to hold compared with signals pulled from live conversations.

Mia: Particle, the company behind the news app Particle News, introduced Radar, a podcast search engine. According to the launch post and co-founder and CEO Sara, Radar searches more than 130,000 actively transcribed podcasts, with about 20,000 new episodes added daily, and Particle describes the material as millions of hours and billions of lines, fully searchable.

Milo: Radar runs on Particle's Podcast Intelligence API, an API and MCP that lets agents search podcast content, and Sara said each episode is fully transcribed, speaker diarized and labeled, with rich entities like people and companies extracted — so the data can establish trends, support deep research, surface podcasts and episodes, or help someone catch up. Radar itself searches transcripts by entity, keyword or key phrase, or semantic meaning, and the maker said an additional search feature is coming soon.

Mia: Sara explained the motivation: in a world increasingly searched by agents, agents cannot hear podcast content unless someone has already transcribed it — which is why the API and MCP had to be built. Over a year ago the team also started bringing the best, most relevant podcast clips into Particle News stories to give each story a human layer of commentary and discussion, and the maker says users told the company they love it.

Milo: Community reaction splits between value and open questions. One commenter said making podcast conversations searchable could make podcast research much less time-consuming, while another said they have many podcasts worth searching but raised questions about what exactly the tool returns.

Milo: StackScope is a new Product Hunt launch billed as technographics for the new web, built by one person, Jonathan, to reveal what newly launched sites are built with — like which sites started using Stripe, Shopify, or Next.js that week. He positions it directly against BuiltWith and Wappalyzer, arguing those tools serve the established web, but by the time a site shows up there it isn't new anymore.

Mia: The freshness comes from the discovery pipeline. Instead of waiting for submissions, it pulls public infrastructure signals to index just over two point two seven million sites and track roughly forty thousand seven hundred technologies. About four hundred thirty-three thousand of those discovered sites get classified as launches — the indie subset — while the rest are ordinary new web, and users can narrow to either group. When StackScope reaches a site before the launch boards, the stated lead is about two weeks early, fourteen point five days in the launch description and fourteen point seven in the on-site stats. The typical launch stack runs around eleven technologies.

Milo: Browsing and search are free, and the paid tiers add Stackdar alerts when new sites pick up a technology, plus API access, CSV export with the published contact details of matched sites, and an MCP server so you can query the data from Claude, ChatGPT, or Cursor. The community reports are concrete — one user said their site was already on StackScope shortly after they started, and another found browsing easy and interesting, though they flagged something about the experience.

Milo: Fotor's Product Hunt launch of Video Agent pitches it to founders, marketers, and creators as a way to build and edit precise motion graphics and video through chat. Fotor's Head of Growth, Coral, says the agent can take an idea, a PDF, or a script and handle the planning, generation, and assembly on a multi-track timeline, auto-orchestrating scenes, timing, visual effects, and kinetic typography. What stands out is that text, numbers, logos, and charts stay editable before rendering, so a typo or a stat change can be fixed without regenerating the whole video. Fotor describes the animated infographics as After Effects-level quality without needing After Effects, with charts that stay live and parametric. Coral says the product came out of an internal experience where a ninety-second AI video had a typo in an animated chart, and correcting one word meant re-running the entire prompt and hoping the next version stayed consistent.

Mia: Coral also frames most AI video tools as giving you a locked MP4. In Fotor's own tests, she says a process that used to take three to five days now takes about forty minutes — a maker-reported result, not independently verified. A commenter who said they worked closely with the team framed the shift as AI-assisted production rather than AI video generation, with one agent orchestrating scripting, scenes, visuals, timing, voice, music, and motion graphics while the user keeps control of the final cut. Community discussion was mostly questions rather than verified results, though one commenter did say the editable multi-track timeline solves a real frustration with AI video tools, and another saw potential in something around the produced output.

Milo: Ask My Wardrobe, launched on Product Hunt by a maker named Bri, is a free AI outfit generator and planner that needs no sign-up to get started. The maker said the original idea was simple: upload a photo of yourself and an item you're considering, and see yourself wearing it before buying. But as people tried it, the maker saw users building different outfits from the same clothes, combining new items with pieces they already owned, saving looks, and returning to compare them — and the outfit planner overtook virtual try-on in popularity, which redirected the product.

Milo: The product now lets users organize the clothes they own, build outfits on a visual editor, plan what to wear, see those plans in a visual calendar, try on clothes virtually before buying, and share outfits with friends for feedback. It's free with no account required to start; the maker says a few free tries are available, more unlock after signing up, and saved outfits stay accessible with an account. Every generated outfit is saved automatically into a personal lookbook, users can sync planned looks with their calendar app, and the maker claims photos stay private unless you choose to share them. On the virtual try-on itself, the maker is candid that the result is directionally right but not perfect, and needs a full-body, forward-facing photo with good lighting and a simple background. Community discussion mostly raised questions rather than confirming results — one commenter said getting the right photo setup was part of the friction.

Mia: Next up, WebTerm Learn, a free, browser-based platform that teaches the terminal as a game. It has twelve courses and a hundred and twenty-nine lessons covering the terminal itself, Git, and Vim, and every command you type gets checked inside a real in-browser sandbox.

Milo: Right, so it's not just watching slides — hand-drawn slides introduce each concept, then you actually run commands and the sandbox confirms whether you got them right. Progress sits on a 3D world map, with a companion that evolves alongside you as you move through the lessons.

Mia: Similarly tool-focused, Orato is an iPhone speech-practice app from solo developer Petros Tepoyan, positioned as a private speech coach. You pick a drill, speak for thirty to ninety seconds, and get scored on pacing, fluency, vocabulary, and coherence. The feedback is granular — every filler word and long pause shows up on a transcript you can read back, and tapping a line reveals what the coach flagged at that moment.

Milo: There are six modes — Impromptu, Defend a Topic, Three Questions, Tell a Story, Elevator Pitch, and Read Aloud — plus a habit layer with a daily reminder, a visible streak, and a two-week progress sparkline. What distinguishes it, per the maker, is privacy: he built it because other coaching apps wanted to upload his voice, whereas on an iPhone with Apple Intelligence the entire pipeline — transcription, scoring, coaching — runs on-device, with no account, and works in airplane mode. His claim is that audio never leaves the phone under any path or setting.

Mia: Two optional paths can push transcript text off-device, though both require sign-in first. iPhones too old for Apple Intelligence can opt into a cloud coach that reads the transcript and asks for Sign in with Apple, and a Pro enhanced-AI-coaching toggle, off by default, sends that text to OpenAI for a deeper report. Practice history stays on the device either way, and pricing is straightforward: free gives you three coached sessions per seven-day week, spendable across any of the drills.

Mia: So two very different looks at cutting out busywork today: Revolte wants teams to stop juggling Cursor, Copilot, and Devin as separate tools, and BrandJet turns public buying signals into signed contracts from a single inbox.

Milo: And with that, thanks for joining us. We'll catch you next time.