AI & Machine Learning
decidi.ai
Audits AI-generated content using a rival model to identify flaws and provide a verdict.
ENDPOINT 1
https://decidi.ai/api/mcp
MCP server metadata
- Name
- decidi
- Version
- 1.0.0
Decidi puts real decisions past more than one frontier model. decidi_challenge is free — use it to have an answer or draft checked by a model from a rival lab before anyone relies on it. decidi_council and decidi_consult are metered to a Decidi account: create an API key at https://decidi.ai/account and send it as an "Authorization: Bearer" header on this connection. Docs: https://decidi.ai/docs.
Known tools 1
decidi_challengeAudit an AI answer or draft with a model from a rival lab — never the one that wrote it — and get back the flaws that matter (invented facts, unsupported claims, missing caveats) plus a one-line verdict.
Inferred read-onlyCONNECT WITH APPROVAL
Client installation
Review this server and its permissions before adding it. Secret placeholders must be set locally.
Codex
~/.codex/config.toml
[mcp_servers.decidi]
url = "https://decidi.ai/api/mcp"
enabled = true
Claude Code
.mcp.json
{
"mcpServers": {
"decidi": {
"type": "http",
"url": "https://decidi.ai/api/mcp"
}
}
}
Claude Desktop
Settings → Connectors → Add custom connector
Name: decidi
Remote MCP URL: https://decidi.ai/api/mcp
Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.
Cursor
.cursor/mcp.json
{
"mcpServers": {
"decidi": {
"url": "https://decidi.ai/api/mcp"
}
}
}
Visual Studio Code
.vscode/mcp.json
Add to Visual Studio Code{
"servers": {
"decidi": {
"type": "http",
"url": "https://decidi.ai/api/mcp"
}
}
}
Generic MCP
Client-specific MCP configuration
{
"name": "decidi",
"transport": "streamable-http",
"url": "https://decidi.ai/api/mcp"
}
MCP Inspector
Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.
TRUST AND VERIFICATION EVIDENCE
Loading Trust v2 evidence…
Checking the associated registrable domain. The BuiltWith key remains server-side.
Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.