Who it is for
People who re-explain the same preferences to every AI tool they use.
A shared record of preferences and judgment calls across agents
Overview
Soup.net records taste and judgment: the preferences and decisions that otherwise get re-explained every session, to every tool.
The mechanism is unusual and worth stating precisely. When an agent faces a judgment call, it checks its hypothesis about the preference, framed as a role, a goal, a preferred option and a reason, and supplies evidence for it. Soup.net returns related prior decisions and records the check itself, so each agent's work informs the next. The design keeps the judgment with the person; what is stored is the record of it. Community tier, documentation published.
People who re-explain the same preferences to every AI tool they use.
Credentials gate the endpoint, so six is the listing's count. The publisher describes one core action, the recipe check, with the rest supporting it.
Availability
SoupNet is listed in the Claude connectors directory only. Andy Forest has no app under the same domain in the ChatGPT directory; 392 of the 3,150 products here are on both marketplaces, and this is not one of them.
The MCP server behind a Claude connector is usually the same server a ChatGPT app would need, so the absence is a distribution decision rather than a technical one.
Claude listing
SoupNet publishes 6 tools to Claude. That is a mid-sized surface: 28.1% of the 1,253 Claude-listed connectors expose fewer.
The median Claude-listed connector filed under Productivity publishes 12, so this one sits 6 tools below its bucket's median, 314th of 462.
It falls in the 4-6 tools band, which holds 16.2% of the Claude directory.
SoupNet publishes a remote MCP endpoint. Claude connects to it over HTTP; there is nothing to install locally.
Every connector in the Claude directory is remote, all 1,253 of them, so transport is not a differentiator there.
https://mcp.soup.net/mcp SoupNet is community-listed in the Claude directory, the tier 770 of 1,253 connectors sit in (61.5%). It is the default for a self-published server, and says nothing either way about quality.
In the publisher's words
Soup.net gives your AI agents a persistent, shared record of your taste and judgment — the preferences and decisions you'd otherwise re-explain every session, to every tool. Agents use one core action: the recipe check. When your agent faces a judgment call, it checks its hypothesis about your preference — "As a [role] working on [goal], I prefer X so that Y" — with evidence. Soup.net returns your related prior decisions, and the check itself is recorded, so every agent you use gets smarter about you from every other agent's work. Your judgment stays yours; Soup.net is where it gets recorded, and the check log is where you watch it being exercised on your behalf. Because Soup.net connects through open standards, the same corpus serves Claude, your coding agents, and any other AI tool you work with — one set of recipe books across vendors, not a separate memory per product. Recipe books can be shared with collaborators, so a team's accumulated judgment compounds instead of living in one person's chat history. Connect in one click: sign in, choose which recipe books to share and whether each is read-only or writable, and your agent receives a briefing with your corpus context. Free hosted service; MIT-licensed and self-hostable.
Quoted from the connector's own Claude directory listing as captured in August 2026. Node8 did not write it and does not vouch for it.
Nearby
Portable memory shared across connected AI apps
ChatGPT
A versioned markdown context layer shared by people and agents
Claude
Signed, citation-backed memory recall for AI sessions
ChatGPT
A consented memory layer shared across separate AI assistants
Claude
Scheduling configuration inspected and edited through chat
ChatGPT
Venue CRM for inquiries, quotes, and event orders
ChatGPT
This is an independent catalogue entry compiled by Node8 from SoupNet's public marketplace listing as it stood in August 2026. Listing facts come from the marketplaces; the comparisons against the rest of the directory and any live-endpoint measurements are Node8's. Node8 is not affiliated with Andy Forest, and inclusion here is not an endorsement.
Node8 builds MCP servers end to end: deciding which tools are worth exposing, the authentication and permission model, the review submissions, and the onboarding that gets them used. One server serves ChatGPT, Claude, and Microsoft Copilot.