Second Read runs on your own machine — the client starts it, so there is no endpoint to ping. 190 installs a week from pypi. Last commit 29 Oct 2026.
Reviews drafts: fact-checks claims against web sources and builds the strongest counter-case.
Nothing serious here today
Today is the operative word: we check Second Read every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
Connect this server
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add second-read -e 'ANTHROPIC_API_KEY=<ANTHROPIC_API_KEY>' -e 'TAVILY_API_KEY=<TAVILY_API_KEY>' -- uvx second-read{
"mcpServers": {
"second-read": {
"command": "uvx",
"args": [
"second-read"
],
"env": {
"ANTHROPIC_API_KEY": "<ANTHROPIC_API_KEY>",
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}[mcp_servers.second-read]
command = "uvx"
args = ["second-read"]
env = { ANTHROPIC_API_KEY = "<ANTHROPIC_API_KEY>", TAVILY_API_KEY = "<TAVILY_API_KEY>" }{
"mcpServers": {
"second-read": {
"command": "uvx",
"args": [
"second-read"
],
"env": {
"ANTHROPIC_API_KEY": "<ANTHROPIC_API_KEY>",
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}{
"servers": {
"second-read": {
"type": "stdio",
"command": "uvx",
"args": [
"second-read"
],
"env": {
"ANTHROPIC_API_KEY": "<ANTHROPIC_API_KEY>",
"TAVILY_API_KEY": "<TAVILY_API_KEY>"
}
}
}
}Put your own values in place of
<ANTHROPIC_API_KEY> (Anthropic API key, used for the review's model calls), <TAVILY_API_KEY> (Tavily API key, used for web search).
The author’s registry entry marks them as required.
Alternatives to Second Read
same job, measured the same wayFact-check claims against live web sources: verdict, consensus, cited sources, PII redaction.
Checks the most important factual claims in AI output against live web sources by default.
Real-time fact-checking MCP. Verifies claims against live sources with structured verdicts.
Hands bulk reading to cheap OpenRouter models, masks secrets, checks each answer against the source
Check the factual claims in a draft or an answer: a quick verdict, or a deep check with sources.
Checks the context files a coding agent reads against the git index, and reports what drifted.
Fact-check claims against live web sources: verbatim evidence, coherence-gated verdict.
Ask for a model and Facet does the rest: it builds, checks and serves a model for one task.
Second Read — questions
Answers built from our own checks of this server.