Keydris, Inc.
Keydris provides authorization infrastructure for AI agents. It lets organizations assign bounded authority through policies and evaluate each governed action before execution, while recording the authorization decision as evidence.
At a Glance
- Businesses and organizations deploying AI agents
- Platform engineering and security teams
- Enterprise governance, risk, and compliance teams
- Insurance and other industries where agents take consequential actions
- +1 more
AI Tools by Keydris, Inc.
(1)
Keydris
AI Agent Authorization Layer
Discussions
No discussions yet
Be the first to start a discussion about Keydris, Inc.
Latest News
keydris-cli update added the guided `keydris run -- claude` path so Claude's own egress is governed.
Keydris published an audited open-source snapshot of keydris-cli under Apache License 2.0.
Founder Ahmed Isse published 'Software is beginning to act,' the Keydris thesis on agent authority.
Keydris published 'How Authority Verification Works for Multi-Agent AI in Insurance.'
Products & Services
A control plane for defining policies, assigning bounded authority to agents, issuing short-lived Keydris Identity Tokens (KITs), verifying authority at action time, supporting approval and revocation, and recording decisions.
A native CLI and local broker for connecting Claude Code and OpenAI Codex sessions to Keydris. It supports initialization, session identity, proxy-based enforcement, local command gating, status, and revocation; distribution is provided through npm and standalone installers for macOS, Linux, and Windows/WSL2.
Open-source, Apache-2.0 Node.js and Python libraries that let an MCP server verify a KIT and redeem per-action credentials through a Keydris gateway, so the MCP server does not need a credential of its own.
Market Position
Keydris positions itself as action-time authorization infrastructure for AI agents rather than an identity provider, general IAM replacement, MCP server, model runtime, or network firewall. Its differentiation is the combination of policy-defined authority, short-lived verifiable authority references, per-action checks at an execution boundary, credential isolation, revocation, and auditable decision evidence. The adjacent competitive set includes IAM/OAuth and policy-engine products, MCP authorization approaches, and AI-agent security/governance platforms; Keydris says it complements rather than replaces those systems.
Leadership
Founders
Ahmed Isse
Founder and CEO of Keydris; his public profile describes experience across e-commerce, HR technology, and fintech, including co-founding Dibsy and previously co-founding 9to5.ai.
Osman Isse
Co-founder and COO of Keydris; his public LinkedIn profile identifies Northumbria University as part of his background.
Abdulla Adam M
Co-founder and CTO of Keydris; his public profile describes him as a software engineer with a previous medical background and experience in software engineering.
Executive Team
Ahmed Isse
Founder & CEO
Previously co-founded Dibsy and 9to5.ai; public profile describes work across e-commerce, HR tech, and fintech.
Osman Isse
Co-founder & COO
Public profile identifies Northumbria University in his background.
Founding Story
Keydris was started around the premise that authentication and access do not answer the action-level question created by autonomous software: whether a particular agent is authorized to perform a particular action against a particular resource right now. Its initial vision is to provide policy-defined, revocable authority and a decision record at the boundary before an agent action executes.
Business Model
Revenue Model
Subscription software for business and organizational users, with self-serve monthly plans based on agents, policies, team members, and KIT issuance capacity; verification is unlimited. Enterprise customers use custom commercial terms.
Pricing Tiers
1 agent, 1 policy, 1 team member, 1,500 KIT issuances per month, unlimited verification, and no credit card required.
3 agents, 3 policies, 1 team member, 5,000 KIT issuances per month, and unlimited verification.
10 agents, 10 policies, 5 team members, 25,000 KIT issuances per month, unlimited verification, and team administration with roles.
Custom scale, organizational governance, enterprise administration and roles, and custom commercial terms.
Target Markets
- Businesses and organizations deploying AI agents
- Platform engineering and security teams
- Enterprise governance, risk, and compliance teams
- Insurance and other industries where agents take consequential actions
- Developers building MCP servers, agent harnesses, and tool integrations
- Governing AI coding agents before repository reads, pull requests, merges, and deployments
- Controlling MCP tool calls and resource reads
- Protecting credential-free MCP servers with per-request authority redemption
- Governing Slack actions and selected channel reads/posts
- Checking local shell commands invoked through supported agent harnesses
- Multi-agent enterprise workflows such as insurance intake, risk, eligibility, and fraud processing