fdkey is an AI agent verification middleware designed for use with Model Context Protocol (MCP) servers and HTTPS APIs. Unlike traditional CAPTCHAs that distinguish humans from bots, fdkey operates as an inverse CAPTCHA, admitting capable large language models (LLMs) while rejecting humans, scripts, and less advanced models. This approach enables platforms to restrict access to AI agents that demonstrate sufficient reasoning abilities, ensuring that only qualified LLMs can interact with protected endpoints.
The tool offers a range of verification puzzle types tailored for LLMs, including semantic multiple-choice questions over near-synonyms, contradiction detection, semantic ranking, rule induction, multi-constraint satisfaction, and mapping rare-language concepts across cultures. Each verification attempt increases the difficulty for subsequent agents, making it progressively harder to bypass the system. The middleware is intended for scenarios where verifying the caller as a capable AI agent is valuable, such as agent-only platforms, data services for AI consumption, agent-to-agent marketplaces, gating access to sensitive or dangerous tools, serving denser payloads to verified agents, and enabling differentiated pricing or payment rails for agents versus humans.
fdkey is delivered as middleware for MCP servers and as an SDK for HTTPS APIs. It features a one-line installation process and allows developers to designate which tools or endpoints require gating. The service is open source, distributed under the MIT license, and supports multi-region routing. A live API is available, and users can obtain an API key for integration. The platform provides a free tier supporting up to 3,000 verifications per month, with higher usage limits available upon request.
As an open-source solution, fdkey is positioned as agent verification infrastructure for the emerging agent economy, supporting use cases ranging from agent social networks and simulation environments to secure marketplaces and specialized data APIs. Its capability gating mechanism serves as both an access control and a quality filter, enabling operators to admit only frontier models if desired.
In the Infrastructure & Backend space, fdkey takes a focused approach. It focuses on securing MCP servers by requiring AI agents to solve LLM-based puzzles before access is granted. It is built as an open-source project for ai infrastructure developers. fdkey is open source under the MIT license. The product ships for the web, the command line, and API, and it can be self-hosted.
fdkey first shipped in 2026. Development happens publicly on GitHub with 27 commits in the last 90 days. Key capabilities include LLM verification, MCP middleware, and anti-bot puzzles. It exposes integrations via an MCP server.
Latest indexed changes and source events
Other apps tracked under the same category.
