Knocklog MCP server
Commit Nov 9, 2026Analyze how websites interact with AI agents and verify the authenticity of visits from ChatGPT, Claude, or Perplexity.
Also known as knocklog/api, api
- Stars
- 0
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 1.0.1
- Last commit
- Nov 9, 2026
- License
- MIT
What is Knocklog MCP?
Knocklog analyzes how websites interact with AI agents and verify the authenticity of visits from ChatGPT, Claude, or Perplexity.
Summary generated from the project's README
About Knocklog MCP server
Knocklog provides analytics specifically designed to track how AI agents interact with websites. It allows developers and site owners to identify which agents visit their domains, verify if those agents are authentic, and monitor the content they read. The service also tracks visitors and leads that originate from AI assistants.
Users can interact with the service through a standard REST API or via the Model Context Protocol. The API requires different types of credentials depending on the action: an ingest key for sending traffic data and an admin key for reading reports. The MCP implementation offers three distinct modes, including a public directory with no sign-in required and a full access mode where keys are passed as arguments. Tools like scan, classify, and traffic allow for deep inspection of agent behavior and site compatibility.
What you can do with it
- 1Verify agent authenticityIdentify and confirm if a specific web request is actually coming from a legitimate AI bot.
- 2Audit site agent compatibilityCheck how a specific domain currently handles or treats incoming AI agents.
- 3Monitor AI driven trafficRetrieve detailed reports regarding the volume and nature of traffic generated by AI assistants.
Written from the project's README. Every use case is checked against a quote from it.
How to install Knocklog MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Knocklog MCP in Claude Code
$ claude mcp add --transport http api https://api.knocklog.ai/mcp/directoryInstall Knocklog MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"api": {
"url": "https://api.knocklog.ai/mcp/directory"
}
}
}Install Knocklog MCP in VS Code
.vscode/mcp.json
{
"servers": {
"api": {
"type": "http",
"url": "https://api.knocklog.ai/mcp/directory"
}
}
}Install Knocklog MCP in VS Code CLI
$ code --add-mcp '{"name":"api","type":"http","url":"https://api.knocklog.ai/mcp/directory"}'Install Knocklog MCP in Codex
$ codex mcp add api --url https://api.knocklog.ai/mcp/directoryInstall Knocklog MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.api] url = "https://api.knocklog.ai/mcp/directory"
Install Knocklog MCP in Gemini CLI
$ gemini mcp add -s user --transport http api https://api.knocklog.ai/mcp/directoryInstall Knocklog MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"api": {
"url": "https://api.knocklog.ai/mcp/directory",
"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 reachableknocklog/api is public and reachablegithub.com/knocklog/api ↗Pass
- Package names this repository as its sourceNot checked
- Build provenance (Sigstore attestation)Not checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-09github.com/knocklog/api ↗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 71 msPass
README
View on GitHubKnocklog API
The public description of the Knocklog API: analytics for AI agents. See which AI agents visit a website, whether each one is who it claims to be, what they read, and the visitors and leads AI assistants send.
- Docs: https://knocklog.ai/docs/api
- OpenAPI 3.1:
openapi.json, also served live at https://api.knocklog.ai/openapi.json - Base URL:
https://api.knocklog.ai
Quick start
# Check how a site treats AI agents (no key)
curl "https://api.knocklog.ai/v1/scan?domain=example.com"
# Identify one request
curl -X POST "https://api.knocklog.ai/v1/classify" -H "Content-Type: application/json" \
-d '{"user_agent": "Mozilla/5.0; compatible; GPTBot/1.4; +https://openai.com/gptbot", "ip": "20.171.206.10"}'
# Create a project: returns an ingest key, an admin key and install code
curl -X POST "https://api.knocklog.ai/v1/projects" -H "Content-Type: application/json" -d '{"domain": "example.com"}'
# Read the report
curl "https://api.knocklog.ai/v1/traffic?days=30" -H "Authorization: Bearer $KNOCKLOG_ADMIN_KEY"
The ingest key (kl_ing_…) only sends traffic and is safe in a collector on the site. The admin key (kl_adm_…) reads reports; keep it private. More in examples/.
Endpoints
| Endpoint | What it does | Key |
|---|---|---|
GET /v1/agents |
List known AI agents | none |
POST /v1/classify |
Identify and verify one request | none |
GET /v1/scan |
Check how a site treats AI agents | none |
POST /v1/scan |
Check how a site treats AI agents (JSON body) | none |
GET /v1/plans |
List plans | none |
POST /v1/projects |
Create a project | none |
POST /v1/events |
Send requests from a collector | ingest (optional) |
POST /v1/logs |
Import access logs | ingest |
GET /v1/plan |
Plan and usage | ingest or admin |
GET /v1/traffic |
Traffic report | admin |
GET /v1/account |
Account overview | admin |
GET /v1/ready-files |
Ready-made robots.txt and llms.txt (Pro) | admin |
POST /v1/keys/ingest |
New ingest key | admin |
POST /v1/project/summary |
Weekly summary by email | admin |
POST /v1/integrations |
Webhook, Slack and company emails | admin |
POST /v1/install/cloudflare |
Install the collector on Cloudflare | admin |
POST /v1/billing/checkout |
Checkout link for a paid plan | admin |
POST /v1/billing/portal |
Link to manage the subscription | admin |
GET /v1/badge/{project}.json |
Badge status | none |
GET /v1/badge/{project}.svg |
Badge image | none |
Webhooks: company.visited and lead.from_ai, signed with x-knocklog-signature (example).
For AI agents
The same features over the Model Context Protocol:
https://api.knocklog.ai/mcp/directory: public tools, no sign-inhttps://api.knocklog.ai/mcp/account: one project, OAuth sign-inhttps://api.knocklog.ai/mcp: every tool, keys as arguments
Listed in the official MCP Registry as io.github.leonlesenechal/knocklog.
Also from Knocklog
- knocklog/ai-agents: open directory of AI agents and crawlers (JSON, CSV)
Support
Questions or a missing endpoint: open an issue or write to contact@knocklog.ai. Using the API is subject to the terms. The files in this repository are under the MIT license.
Knocklog MCP questions
How do I add Knocklog MCP to Claude Code?
Run claude mcp add --transport http api https://api.knocklog.ai/mcp/directory in your terminal. Add --scope user to make it available in every project.
How do I add Knocklog MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
How can I check how a website treats AI agents?
You can use the scan endpoint without an API key to check a specific domain.
What is the difference between an ingest key and an admin key?
The ingest key is used to send traffic data and is safe for use in site collectors, while the admin key is used to read reports and must be kept private.
Can I identify if a specific request is from a real AI agent?
Yes, you can use the classify endpoint to identify and verify a request using a user agent and IP address.