Fify MCP server
Commit Aug 25, 2026Transforms grounded AI responses into interactive comparisons, plans, checklists, timelines, and decision views.
Also known as renfei-design/fify-plugin, fify-plugin
- Stars
- 1
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 0.1.0
- Last commit
- Aug 25, 2026
- License
- Apache-2.0
What is Fify MCP?
Fify transforms grounded AI responses into interactive comparisons, plans, checklists, timelines, and decision views.
Summary generated from the project's README
About Fify MCP server
Fify transforms grounded AI responses into structured interactive formats like timelines, checklists, plans, and decision views. It works by taking plain-language answers provided by a host like Codex and rendering them through a deterministic information-view compiler. The system preserves the original text as an authoritative fallback and performs no consequential actions on its own.
Developers can install the plugin directly in Codex using the marketplace commands or connect to it via a public Streamable HTTP MCP endpoint for use in ChatGPT or other compatible MCP clients. Using the remote endpoint in ChatGPT requires enabling developer mode in your security settings. The local version runs on Node.js 20 or newer and does not require a Fify account or any external API keys. The bundled server imposes limits of twenty successful renders per UTC day and two concurrent renders per host identity.
What you can do with it
- 1Create interactive project timelinesAsk the AI to display specific milestones as a visual timeline.
- 2Generate actionable rollout checklistsConvert a text-based rollout plan into a checklist for task management.
- 3Perform comparative analysisRequest a comparison of different options based on specific metrics like cost and risk.
Written from the project's README. Every use case is checked against a quote from it.
How to install Fify MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Fify MCP in Claude Code
$ claude mcp add --transport http fify-plugin https://fify-chatgpt.renfei1992.chatgpt.site/api/mcpInstall Fify MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"fify-plugin": {
"url": "https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp"
}
}
}Install Fify MCP in VS Code
.vscode/mcp.json
{
"servers": {
"fify-plugin": {
"type": "http",
"url": "https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp"
}
}
}Install Fify MCP in VS Code CLI
$ code --add-mcp '{"name":"fify-plugin","type":"http","url":"https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp"}'Install Fify MCP in Codex
$ codex mcp add fify-plugin --url https://fify-chatgpt.renfei1992.chatgpt.site/api/mcpInstall Fify MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.fify-plugin] url = "https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp"
Install Fify MCP in Gemini CLI
$ gemini mcp add -s user --transport http fify-plugin https://fify-chatgpt.renfei1992.chatgpt.site/api/mcpInstall Fify MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"fify-plugin": {
"url": "https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp",
"type": "http"
}
}
}Automated checks
Ran Nov 15, 2026Full report →Not verified yet. A server needs a score of 80, no serious finding, and enough data for us to check. How checks work · What to fix for more points
- Source repository linked and reachablerenfei-design/fify-plugin is public and reachablegithub.com/renfei-design/fify-plugin ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-08-25github.com/renfei-design/fify-plugin ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNot checked
- No critical advisory in the package or its dependenciesNot checked
- No hidden instructions to the model0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNot checked
- Remote endpoint answers the MCP handshakeinitialize answered with protocol 2025-06-18 in 1158 msPass
README
View on GitHubFify plugin marketplace
Fify turns grounded AI answers into trusted interactive comparisons, plans, checklists, timelines, and decision views. It preserves the complete plain-language answer as an authoritative fallback and performs no consequential actions.

This repository is a Git-backed marketplace for Codex. The plugin is self-contained: it requires Node.js 20 or newer, performs no separate npm or pnpm install, asks users for no Fify account or end-user API key, and runs a bundled deterministic information-view compiler locally.
Install in Codex
codex plugin marketplace add renfei-design/fify-plugin
codex plugin add fify@fify
Start a new Codex task after installation, then ask normally or tag @Fify:
- Compare these options by cost and risk.
- Make this rollout plan a checklist I can work through.
- Show these milestones as a timeline.
The repository is the marketplace. No package installation, Fify account, or end-user API key is required.
Use with ChatGPT
This Git marketplace cannot be added directly as a user-defined ChatGPT marketplace. You can still use Fify in ChatGPT through a developer-mode MCP connection:
- Open Settings → Security and login and enable Developer mode.
- Open the ChatGPT Plugins page and select the plus button.
- Name the connection Fify and use the public MCP endpoint below.
- Review the discovered read-only tool, create the connection, and add it to a new chat.
https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp
Developer mode availability depends on the user's account and workspace policy. Public discovery across ChatGPT and Codex requires publication to OpenAI's universal plugin directory. This repository is the Codex distribution source and the hosted endpoint is the direct ChatGPT connection source; neither makes Fify automatically discoverable in ChatGPT. See OpenAI's connection guide and plugin packaging guide.
Use from another MCP client
Fify also exposes a public Streamable HTTP MCP endpoint for compatible clients:
https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp
The tool always returns a complete plain-language fallback. The inline interactive view requires a host that supports the MCP Apps UI extension.
See MARKETPLACE_LISTING.md for verified listing metadata, capabilities, limitations, and submission copy.
What is included
- A Codex plugin manifest and Fify information-UI skill.
- A bundled local MCP server with
render_information_ui. - Product screenshots, activation evals, and reviewer cases.
- Validation and end-to-end protocol smoke tests.
- Official MCP Registry metadata for the public remote endpoint.
The local compiler does not research or verify facts. Codex supplies the grounded answer and sources, and Fify validates and renders that material. The bundled server limits each host identity to 20 successful renders per UTC day and two concurrent renders.
Verify locally
npm run check
See BUILD_PROVENANCE.md for the public bundle boundary and SECURITY.md for reporting guidance.
License
Apache-2.0. Bundled dependency notices are in THIRD_PARTY_NOTICES.md.
Fify MCP questions
How do I add Fify MCP to Claude Code?
Run claude mcp add --transport http fify-plugin https://fify-chatgpt.renfei1992.chatgpt.site/api/mcp in your terminal. Add --scope user to make it available in every project.
How do I add Fify MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
Does Fify require an account or API key?
No, Fify is self-contained and does not require a Fify account or an end-user API key.
How can I use Fify with ChatGPT?
You can use Fify in ChatGPT by enabling Developer mode in your settings and connecting to the provided public MCP endpoint.
Does Fify verify the facts in the AI answers?
No, the local compiler does not research or verify facts; it relies on the host to supply grounded answers and sources.
Are there any usage limits?
Yes, the bundled server limits each host identity to 20 successful renders per UTC day and two concurrent renders.