# Whet > Whet watches the feeds, threads, blogs and channels you pick, runs each new item through a managed LLM that scores and summarizes what moved, and lands the daily signal in front of you before you log in. Web-first, self-serve, free to start. Whet is a web research workbench for people who monitor many information sources every day. Creators, founders, researchers, indie operators and small agency leads point Whet at the public sources they care about (Reddit, YouTube, Hacker News, RSS feeds, any public web URL) and Whet does the daily skim for them: ingestion, scheduling, LLM scoring, summarization, daily digest. The product runs in the browser on Whet's managed infrastructure with Gemini Flash-Lite as the default LLM. There is nothing to install, no API key to configure, no CLI to learn. A free tier covers two active sources and 100 LLM analyses per month, enough to try the daily skim on a real workflow without paying or signing a card. ## What Whet does Three beats, in plain English: - **Listen.** Whet keeps an eye on the public conversations that matter to you. Feeds, threads, blogs, channels you pick. Signals you would miss otherwise land in your inbox, decoded into what actually moved. - **Analyze.** Every new item runs through an LLM that scores it against your interests, summarizes the substance, and flags the items worth your time. You read the digest, not the firehose. - **Publish.** Pipe the signal into a draft, export clean markdown, post to your CMS, or fire a webhook downstream. Every artifact is portable, every action is auditable. Whet does not lock you in. ## Sources it supports out of the box Five sources, each backed by an official public API or an open protocol. Whet runs the schedule, manages the rate limits, and ingests on your behalf: - **Reddit.** Reddit API (OAuth, official). Posts, comments and scores from any subreddit you watch. - **YouTube.** YouTube Data API (official). Channels, titles, descriptions, comments, plus full transcript ingestion when captions are available. - **Hacker News.** Algolia HN API (public). Front page, new, jobs, ASKs and the comment threads underneath. Zero auth needed. - **RSS / Atom.** Open protocol, any feed. Blogs, newsletters, podcast notes, changelogs, GitHub releases. Anything that publishes. - **Public web.** Any HTTPS URL. Paste a URL and Whet pulls a clean markdown extract, one-off or on a recurring diff schedule. Plus BYO content via direct upload. Free supports two active sources at once; Basic five; Pro fifteen on web (unlimited on the Desktop add-on). ## Tiers Web-first, self-serve, four tiers: - **Free ($0 forever).** 2 active sources, 100 LLM analyses per month, daily schedule, markdown export. Personal use. No card. - **Basic ($9 / month).** 5 active sources, 1,000 LLM analyses per month, email alerts on matches, commercial-use license, priority signal scoring, all export formats. 7-day free trial, no card. - **Pro ($29 / month).** 15 active sources on web with 5,000 LLM analyses per month, plus the desktop add-on (macOS, Windows, Linux) where BYOK or BYOA lifts every cap. Native Notion, Linear and Cal exports. Priority support. - **Team (contact).** Shared workspaces, role-based access, audit logging, consolidated billing. Opens formally after Pro has traction; for now, talk to us and we set up a Team trial. Whet absorbs the LLM cost on every web tier (Gemini Flash-Lite on Free and Basic, Flash on Pro web) so users can start without configuring keys. The Desktop tier flips this: the user brings their own key or agent CLI and the cap goes away. ## Desktop add-on (Pro and above) The Desktop app is included with Pro. Same workbench, same data, same auth as the web app, with capabilities only a local binary offers: - **BYOK.** Bring your own LLM key (Gemini, Claude, OpenAI). No managed cap. - **BYOA.** Delegate to a local agent CLI (Claude Code, Cursor, Codex). - **Local filesystem.** Read and write to local paths directly from the workbench. - **MCP server.** Local Streamable HTTP server, bearer auth, compatible with Claude Desktop, Cursor, Codex and Cline. - **CLI.** `pnpm install -g @whet/cli`. Same auth as the web workbench. The Desktop add-on is positioned as a natural upgrade for power users who hit the web caps or who want their LLM workload to stay on their own machine. It is not the default surface. ## Who Whet is for Power users solos and small teams who monitor many information sources every day and would rather spend their time on the signal than on the skim. Creators, founders, researchers, indie operators, small agency leads. People who already know feed readers, scraping tools and workflow builders, and who can tell the difference between a real product and AI marketing. Not for casual readers who want a generic news app. Not for teams looking for a templates library or a content calendar. Not for anyone who needs an autoposter; every published item requires explicit confirmation, whether triggered from the workbench, the CLI or an MCP tool call. ## Engagement model Self-serve from the public landing. Sign up free in the browser, add a source, let Whet run a day, see the first digest. Upgrade to Basic or Pro from inside the app when ready. Team and custom integrations go through a 15-minute call with the founder. - Sign up: https://whet.so (Free tier, no card) - Book a call (Team, custom, Pro evaluation): https://cal.com/crewtives - Support: hello@whet.so ## Resources - [Homepage](https://whet.so/): hero, sources, agent surface, pricing, FAQ. - [Pricing](https://whet.so/pricing): full catalog and compare table for Free, Basic, Pro and Team. - [Docs](https://docs.whet.so/): platform concepts, source catalog, CLI reference, MCP integration. - [About](https://whet.so/about): origin, beliefs, model. - [Security](https://whet.so/security): architecture commitments and data handling. - [Privacy](https://whet.so/privacy) and [Terms](https://whet.so/terms). - [Book a 15-minute call](https://cal.com/crewtives). - [Crewtives studio](https://crewtives.com): the parent studio behind Whet. ## What Whet is not - Not a feed reader. Reading the firehose is what Whet replaces. - Not a templates library or content calendar. - Not a scheduler or autoposter. Every publish action needs explicit confirmation. - Not a data vendor. Whet does not resell scraped content. The operator owns their source list and their produced artifacts. - Not a copycat engine. Whet's pipelines surface decoded patterns, aggregated metrics and drafts grounded in the operator's voice, not reworded versions of competitor posts. - Not a tool that scrapes private platforms with burnable accounts. Sources are limited to those backed by an official public API or an open protocol; if a source is not in the catalog, it is not in scope.