Simple Product can read web pages now

What shipped

Web page reading. Following on from last week's web_search, Simple Product can now deep-fetch a specific URL and return its main content as clean markdown. Point it at a blog post, documentation page, GitHub README, product page, or any URL — it'll read the full body and use it as context.

Same "build once, available everywhere" property: sp3k chat, MCP (Claude Desktop, Cursor, etc.), the skills engine, and integrations all pick up the new tool automatically. A skill can now compose "search for X → pick the best result → read it in full → synthesize into a doc," which is a much bigger unlock than either tool alone.

Handles most JS-rendered pages via Firecrawl under the hood. 1 credit per page, metered under the Tools bucket right next to search.

What this unlocks

• Sp3k chat: "Read this URL and summarize the pricing model" or "Compare these two pages and tell me the differences."
• Skills: chain search + read to build up context automatically — competitor blog monitoring, changelog watching, industry-signal roundups.
• MCP surfaces: Claude Desktop and Cursor users get the tool automatically the moment they open SP's connector.

What's next

Building on this: deep research — either as a first-class SP tool (long-form reports with citations) or as a "brief + external agent + writeback" workflow where SP is the durable memory across research projects. More soon.