~/council-of-ai/gspc-plugin (main)
Measure AI. Anywhere.
One command.
The GSPC plugin puts Council of AI's signed measurement toolkit inside your AI platform — Claude Desktop, Cursor, ChatGPT, Open WebUI, LibreChat, and anything that speaks MCP. Verification is free forever.
$ install
Pick your platform. Paste. Done.
The MCP server ships as csoai-governance-mcp on npm (v0.1.0), or run it hosted with no install at all — the server speaks MCP over HTTP streamable at https://councilof.ai/mcp.
Claude Desktop
Open Settings → Developer → Edit Config to reveal claude_desktop_config.json, then paste:
{
"mcpServers": {
"csoai-governance": {
"command": "npx",
"args": ["-y", "csoai-governance-mcp"]
}
}
}
Cursor
Create .cursor/mcp.json in your project root (or ~/.cursor/mcp.json for every project), then paste:
{
"mcpServers": {
"csoai-governance": {
"command": "npx",
"args": ["-y", "csoai-governance-mcp"]
}
}
}
Any MCP client — hosted, no install
Point any client that connects over HTTP streamable — ChatGPT connectors, Open WebUI, LibreChat, and anything else that speaks MCP — at the hosted server:
https://councilof.ai/mcp
One command, from your shell
Smoke-test the server directly:
npx -y csoai-governance-mcp
$ tools --list
What you get
Board read
Read published measurement boards live from your AI session.
Evidence search
Search the evidence behind published measurements without leaving the conversation.
Verification
Check signatures and receipts on any measurement. Free forever.
Attestation issuance
Issue signed attestations from your session. Paid via x402.
$ write-path --status
The write path — coming next
Session recording (write path) is in final development. Cards are born observed_uncontrolled, pass a contamination firewall, and require named-human review before influencing any published measurement. Measurement, not certification — for writers too.