Claude Free Plan Update: File Creation, Connectors, and Skills (What’s Actually New)

Claude Free Plan Update: File Creation, Connectors, and Skills (What’s Actually New)

Anthropic just made a meaningful change to Claude’s positioning: more “agent-like” features are now available without paying.

The headline: Claude free users can now create files, use Connectors, and access Skills—capabilities that were previously tied to paid plans.

Primary sources:

• •CNET — “Anthropic Expands Claude's Free Tier With More Features”

https://www.cnet.com/tech/services-and-software/anthropic-expands-claudes-free-tier-with-more-features/

• •MacRumors — “Anthropic's Claude Gets More Free Features…”

https://www.macrumors.com/2026/02/11/anthropic-claude-more-free-features/

• •Anthropic (Claude blog) — “Cowork: Claude Code for the rest of your work” (context on connectors + skills)

https://claude.com/blog/cowork-research-preview

TL;DR

• •File creation is now in the free plan: Claude can generate documents like PowerPoint, Excel, Word, and PDFs from a conversation. (CNET, MacRumors)
• •Connectors are now free (select availability), including Google Workspace mentioned explicitly. (CNET)
• •Skills are now free, including custom skills (repeatable instruction sets) so you don’t have to re-prompt the same workflow every time. (CNET)
• •Anthropic is also pushing toward longer, more usable free sessions with conversation “compaction” (summarizing earlier context). (MacRumors)

What changed in the Claude free plan (and why it matters)

1) File creation: from chat output → real artifacts

The most obvious upgrade is that Claude’s output no longer has to stop at plain text.

According to reporting, free users can now create and save:

• •PowerPoint slides
• •Excel spreadsheets
• •Word documents
• •PDFs

That matters because it moves Claude from “assistant” to “tool you can hand deliverables to.” If you do recurring work (status decks, outreach lists, reporting templates), this is the kind of capability that can genuinely reduce time-to-output.

Sources:

• •https://www.cnet.com/tech/services-and-software/anthropic-expands-claudes-free-tier-with-more-features/
• •https://www.macrumors.com/2026/02/11/anthropic-claude-more-free-features/

2) Connectors: Claude can pull from the apps where work actually lives

Anthropic’s Connectors link Claude to external services so it can reference (or act on) information you already have in your tools.

CNET calls out Google Workspace specifically as part of what’s now available in the free plan.

MacRumors notes a broader list of connector options (examples include Slack, Asana, Zapier, Stripe, Notion, and others) via Anthropic’s connectors page.

Source:

• •https://www.cnet.com/tech/services-and-software/anthropic-expands-claudes-free-tier-with-more-features/

3) Skills: reusable “instructions” for workflows you repeat

Anthropic’s wording varies by surface, but the idea is consistent: Skills let you create a repeatable, packaged set of instructions so Claude behaves consistently.

Practical examples:

• •A “weekly update” skill that outputs your status in a fixed format
• •A “blog draft” skill that enforces your house style and SEO checklist
• •A “meeting notes” skill that always produces action items + owners + deadlines

Sources:

• •https://www.cnet.com/tech/services-and-software/anthropic-expands-claudes-free-tier-with-more-features/
• •https://claude.com/blog/cowork-research-preview

What this means for users choosing between ChatGPT vs Claude

Even if you don’t care about model benchmarks, this is a big deal for day-to-day usage:

• •Many people evaluate chatbots based on workflow features, not raw intelligence.
• •Moving artifacts + connectors + skills into free lowers the “try it seriously” barrier.

If you’re advising a team or picking a default assistant for operations work, you can treat this update as:

• •Claude being more aggressively positioned as a productivity platform, not just a chatbot
• •An attempt to win users who are sensitive to the free-tier experience (especially after recent changes in the broader chatbot market)

Safety note (connectors + file access changes the risk profile)

Once a chatbot can access files and external tools, the primary risk stops being “bad answers” and starts being “bad actions.”

Anthropic explicitly flags this in its Cowork post, noting you should:

• •limit folder access to what’s necessary
• •expect the assistant to ask before significant actions
• •watch out for prompt injection risks when an agent is interacting with web content

Source:

• •https://claude.com/blog/cowork-research-preview

Bottom line

If you only use AI chatbots for quick Q&A, this won’t change much.

But if you want Claude to behave more like a lightweight agent—creating real files, pulling context via connectors, and following reusable skills—this is one of the most consequential free-tier upgrades we’ve seen recently.

If you want, I can also publish a follow-up guide: “Best Claude Connectors for Founders/Operators (and what to avoid).”