Every change Wellknown observed on this MCP server, newest first, with what it was before and what it became. Tool-surface changes carry the definition diff. Nothing here is edited after the fact.
Added "best_for" (12 tools before, 13 now)
The best AI tools for a topic, ranked — Use this when the user asks for the best AI tools for a task, niche or category — "best AI meeting notes tools", "top AI video editors", "what should I use for logo design". Returns our ranked list for the topic: the same ranking our "Best AI tools for X" pages use — topic match, real-user mentions, how often AI engines read each tool, and viability, with a founder-reviewed consensus list as a sanity gate. A #1 is named ONLY when the leader clears that gate; otherwise the answer says no single top pick is warranted. Every tool carries its pricing model, viability score and page. Not for: a si…
Added "check_mcp_server" (11 tools before, 12 now)
What the official registry says about an MCP server — Use this when the user asks about a specific Model Context Protocol server — whether it is listed in the official MCP registry, what endpoint it publishes, how current its listing is, or whether that endpoint responds. Returns the registry record (name, version, publish and update dates, declared remote endpoints, registry status) plus the result of one anonymous request to a registry-published endpoint. Reports observations only. It does NOT assess whether a server is safe, trustworthy, well-built or suitable — we cannot see what a server does with data, and reachability is not safety. Many …
Recently observed → Live
Live → Recently observed
Added "stack_overlap" (10 tools before, 11 now)
Where an AI tool stack overlaps with itself — Use this when the user asks whether the AI tools they pay for duplicate each other — "are we paying twice", "does our stack overlap", "can we consolidate these". Give it the tools they actually pay for. Returns the pairs we classify the same way, with the shared categories and shared capability tags named for each pair so the user can judge the claim themselves. IMPORTANT: this compares how RightAIChoice classifies two products — not their features, their APIs, or what a team does with them. It is a prompt to review a pair, never evidence that either is redundant. It never recommends cancelli…
Recently observed → Live
Certificate changed, valid to 2026-12-23
Live → Recently observed
Recently observed → Live
Live → Recently observed
Certificate recorded, valid to 2026-10-25
Authorization not required
Changed the definition of "category_landscape", "check_tool_status", "compare_tools" and 2 more
⟨84 unchanged words⟩ and weakest ends — each with its own survival score andlast-verifiedthedate.date our record of it was last rebuilt. Categories are our own classification and tools belong ⟨52 unchanged words⟩
⟨81 unchanged words⟩ on a rolling weekly cycle. Every answer stateswhenBOTH
First tool surface recorded: 10 tools (server version 0.1.0)
Showing the latest 14 events. The API returns up to 500 and filters by kind: ?kind=tool_surface_changed
⟨28 unchanged words⟩ BOTH tools side by side (alive/at-risk/shut-down verdict,last-verifiedthe vendor-link probe date AND the separate catalog-rebuild date, pricing reality, market sentiment) plus ⟨70 unchanged words⟩
⟨47 unchanged words⟩ with a one-line description, current health verdict,last-verifiedthe date our record of it was last rebuilt, and pricing model. If the asked-about tool ⟨73 unchanged words⟩
⟨81 unchanged words⟩ Each result carries its own health verdict,last-verifiedthe date our record of it was last rebuilt, and pricing model. Matching is keyword and fuzzy ⟨62 unchanged words⟩