What Happened to thesecondbrain.io?

If you saved a link to thesecondbrain.io some months ago and went back to it recently, you found something unexpected: the site loads fine, your login works, and the whole thing is called Rescroll.

Nothing broke — here is what actually changed, checked against the live site on 4 September 2026.

The short version

thesecondbrain.io still resolves. It was not shut down, sold off, or redirected to a different company. The domain now serves a product called Rescroll, the app still lives at app.thesecondbrain.io, and the underlying tool is recognisably the same one.

What changed is who it is for. The old positioning was general-purpose personal knowledge management — save things, search them, chat with them, in the tradition of Notion, Obsidian and Evernote. The new positioning is a content tool for creators, agencies and social media managers. The headline on the homepage now reads "The content creation cheat code," and the hero output is a ready-to-post carousel.

That is a repositioning, not an exit. Worth being precise about, because "what happened to X" posts usually imply a company died, and this one did not.

What stayed the same

If you were using it as a library, most of what you relied on is still there. You can drop in TikToks, Reels, YouTube, podcasts, PDFs and articles. Semantic search still finds things by meaning. Projects and tags still exist. There is still a Chrome extension for saving as you browse. Chat-with-your-library is still the core interaction.

Pricing is three tiers, credit-metered: Basic at $6/month for 500 credits, Pro at $20 for 2,500, Ultimate at $55 for 10,000. Visual Boards, auto-organisation, premium models and carousel generation sit above the Basic tier.

Your saved content also survives the source disappearing. Their documentation is explicit that if a creator deletes the TikTok or a brand pulls the Reel, your processed copy stays. That is a real property and not every tool in this space has it.

What changed

The output side is where the product has moved.

Carousel generation is now a headline feature — a prompt produces a full deck with generated images per slide, titles, caption, hashtags and a suggested first comment. Image and short-video generation are built into the chat on the paid tiers. There is a Visual Boards canvas for running several AI chat threads beside your sources at once.

The comparison pages tell the same story. They now benchmark against ChatGPT, Notion, Canva, Castmagic, Opus Clip, Taplio and Jasper. Three of those are creator tools and two are social-copy generators. That is a different competitive set than a general knowledge tool would pick.

The thing it does that others do not

Being fair about a competitor means naming what they are genuinely good at, so: frame-by-frame visual analysis of short-form video.

Reels, TikToks and YouTube Shorts get watched, not just transcribed. The system reads on-screen text, registers the cuts, and pairs that with the audio. For long-form YouTube under thirty minutes, detailed visual analysis is an opt-in checkbox at paste time.

If your work involves studying why a piece of short-form video performed, that is a real capability and most libraries do not have it — including ours. They also import up to 50 videos from a YouTube channel in one go, which is a genuinely useful way to map a catalogue quickly.

Whether you should stay

Two honest answers, depending on why you signed up.

If you make content for a living, it probably got better for you. The tool is now built around the workflow you actually have: study what worked, draft in that style, produce the asset. The features added over the last few months are aimed squarely at that, and the general-purpose version was never optimised for it.

If you wanted a durable personal knowledge base, it is pointed elsewhere now. Not broken for that use — the library, search and chat all still work — but the roadmap energy is visibly going into creator output, and the comparison set no longer includes the tools you were weighing it against.

One specific gap matters if you are working with AI assistants: Rescroll has no MCP server and no public API. Your library is reachable from inside their product and not from outside it. If your pattern is asking Claude or ChatGPT questions against your own saved material, or having an agent read and write to your library, that is not available here. For a creator working inside one tool, this may not matter at all. For anyone building an assistant workflow around their archive, it is the whole question.

The wider point

There is a pattern worth noticing, and it is not a criticism of anyone.

A product called "The Second Brain" named a category broad enough to be hard to sell. Everybody needs one in principle; almost nobody wakes up wanting one specifically. Tools built on that promise tend, eventually, to pick a job that people already know they have — and creator workflow is one of the clearest of those jobs, with an obvious willingness to pay.

We are not neutral here. BrainTube is still built on the general promise, and we have chosen a different bet: that the durable, portable, agent-reachable library is worth building even though it is a harder thing to explain. That bet could be wrong. It is at least worth stating plainly rather than pretending we are describing this from outside.

If you are leaving a tool for another one, the question worth asking of whoever you move to is not which features they have today. It is whether you can get everything out again, and whether anything other than their own app can read it.

Where BrainTube stands

We are not describing this from outside, so here is our own position on the same axis. BrainTube is a compiled knowledge base that any MCP client can query: one read-write MCP server covering the whole library, 44 tools listed on the connect page as of 5 September 2026, included on every plan including Free, and export over MCP as a ZIP of Markdown files. Where that lands against Recall specifically is set out in our comparison.

And the half that does not favour us. BrainTube has nothing like Rescroll's frame-by-frame analysis of short-form video, and no image or video generation at all. Social-post capture is not something we have verified, so it is not a capability to plan around. There are no native iOS or Android apps. The in-app JSON export is capped to the 500 most-recent items, and ZIP import has never been tested against a real export from another tool, so we do not offer it as a migration path.

How these figures were checked

Every claim about Rescroll here was confirmed by a scripted read of the rendered page at www.thesecondbrain.io on 4 September 2026 — the live homepage, its pricing section and its FAQ. Pricing, credit allowances, the content types supported, the frame-by-frame analysis behaviour, the 50-video channel import, the deleted-source retention policy and the comparison set are all stated there. One practical note if you go and check: the dollar prices render only with scripts enabled, so a plain fetch of the page returns the credit allowances without them.

The absence of an MCP server or public API is an absence of any mention on their public marketing site as of that date, not a statement from the company. If they have shipped either since, their site is the authority and we will correct this.

Product and positioning change quickly. If you spot something here that has gone stale, tell us and we will correct it with the date attached.