{"$schema":"https://wellknown.network/schemas/agent-record-v1.json","schemaVersion":"1","id":"ag_r8pqr3h4h593","handle":"neuraldefend-mcp","url":"https://wellknown.network/agents/neuraldefend-mcp","links":{"self":"https://wellknown.network/agents/neuraldefend-mcp/record.json","html":"https://wellknown.network/agents/neuraldefend-mcp","markdown":"https://wellknown.network/agents/neuraldefend-mcp/record.md","api":"https://wellknown.network/api/v1/agents/neuraldefend-mcp","status":"https://wellknown.network/api/v1/agents/neuraldefend-mcp/status","claim":"https://wellknown.network/agents/neuraldefend-mcp/claim","claimApi":"https://wellknown.network/api/v1/claims","claimDescriptor":"https://wellknown.network/agents/neuraldefend-mcp/claim.json","badge":"https://wellknown.network/agents/neuraldefend-mcp/badge.svg","openapi":"https://wellknown.network/openapi.json","history":"https://wellknown.network/api/v1/agents/neuraldefend-mcp/history","tools":"https://wellknown.network/api/v1/agents/neuraldefend-mcp/tools"},"ard":{"identifier":"urn:air::server:neuraldefend-mcp","type":"application/mcp-server-card+json"},"kind":"mcp_server","declared":{"name":"neuraldefend-mcp","summary":"Secure MCP server for Neural Defend NeuroVerify — deepfake and AI-generated media detection for MCP agents","description":"# neuraldefend-mcp\n\n`neuraldefend-mcp` 2.0.0 is a secure, local Model Context Protocol (MCP) server for\nNeuroVerify image and video authenticity analysis. It gives an MCP client two tools while\nkeeping API credentials out of tool arguments, restricting which local media can be read,\nand returning a bounded integration contract instead of the Python SDK's raw response.\n\nUse the results as decision-support signals. A low-risk result is not proof that media is\nauthentic, and a high-risk result should trigger human review rather than automatic\nrejection.\n\n## How it works\n\nThe MCP client starts `neuraldefend-mcp` as a child process and communicates with it over\nstandard input and standard output:\n\n```text\nMCP client\n  └─ local stdio (MCP JSON-RPC)\n      └─ neuraldefend-mcp\n          ├─ validates and opens an allowlisted local file once\n          └─ streams that open handle over TLS through the NeuroVerify Python SDK\n              └─ NeuroVerify Detection API\n```\n\nThere is no local HTTP listener. Standard output is reserved for the MCP protocol; safe\nstartup errors are written to standard error. The server does not return media bytes to the\nMCP client and does not load an entire video into memory. It validates the opened file\nhandle and uploads that same bounded handle, reducing path substitution and\ntime-of-check/time-of-use risk.\n\n## Requirements\n\n- Python 3.10 or newer.\n- A NeuralDefend API key authorized for the requested image or video analysis.\n- An MCP client that can launch a local stdio server.\n- Network access to the selected NeuroVerify API environment.\n- Read access to dedicated local directories containing only media authorized for upload.\n\n## Installation\n\nInstall the published 2.0.0 release:\n\n```bash\npython -m pip install \"neuraldefend-mcp==2.0.0\"\n```\n\nThe package installs the `neuraldefend-mcp` command. The equivalent module entry point is:\n\n```bash\npython -m neuraldefend_mcp\n```\n\nBoth entry points run the same stdio-only server. Starting either one direc…","publisher":{"name":"Neural Defend","url":null},"homepage":"https://github.com/Neural-Defend/NeuralDefend-SDKs/tree/main/packages/mcp","repository":"https://github.com/Neural-Defend/NeuralDefend-SDKs/releases","version":"2.0.0","license":"MIT License\n        \n        Copyright (c) 2026 Neural Defend\n …","protocols":["mcp"],"tags":["ai-agent","ai-detection","ai-generated-content","ai-generated-image-detection","ai-generated-video-detection","ai-image-detection","ai-video-detection","authenticity","deepfake","deepfake-detection","deepfake-detection-api","face-authenticity","fraud","fraud-prevention","generative-ai-detection","identity-verification","image-authenticity","llm-tools","manipulation-detection","mcp","media","media-authenticity","media-forensics","model-context-protocol","neural-defend","neuraldefend","neuroverify","security","spoof-detection","synthetic-media-detection","video-authenticity"],"pricing":null,"endpoints":[{"url":"pypi:neuraldefend-mcp","type":"package_pypi","auth":null,"probeable":false}],"skills":null,"tools":null,"extra":null,"attribution":{"kind":"pypi","name":"pypi","license":"pypi","repoUrl":"pypi","summary":"pypi","version":"pypi","description":"pypi","homepageUrl":"pypi","publisherName":"pypi"}},"derived":{"capabilities":[{"slug":"media.video-generation","name":"Video Generation","confidence":0.675,"provenance":"derived"},{"slug":"security.identity","name":"Identity & Access","confidence":0.675,"provenance":"derived"}],"categories":["media","security"],"language":"en"},"observed":{"status":"unknown","statusReason":"Distributed as a package to run locally; no network endpoint to check.","lastOkAt":null,"lastProbedAt":null,"statusComputedAt":null,"reliability30d":null,"latestObservations":[],"tools":null,"package":{"name":"neuraldefend-mcp","registry":"pypi","observedAt":"2026-10-22T15:29:06.298Z","publishedAt":"2026-08-22T05:36:43.491235Z","latestVersion":"2.0.0"},"toolSurface":null,"endpointFacts":[]},"verification":{"claimed":false,"claimedAt":null,"proofs":[]},"provenance":{"sources":[{"source":"pypi","key":"neuraldefend-mcp","url":"https://pypi.org/project/neuraldefend-mcp/","firstSeenAt":"2026-09-10T08:24:59.877Z","fetchedAt":"2026-10-22T15:27:11.896Z","normalizedAt":"2026-10-22T15:27:11.896Z"}]},"firstSeenAt":"2026-09-10T08:24:59.877Z","updatedAt":"2026-10-22T15:29:06.298Z"}