About
Built because autonomous action needs verified intent.
Intended exists because agents should not be able to change reality without verification. What they are about to do must match what was intended, authority must be granted cryptographically, and the result must be provable after the fact.
Company
Intended, Inc.
Austin, Texas, USA.
We are building the category of intent verification infrastructure for autonomous agents. The job is simple to state and hard to get right: verify intent before execution, grant authority cryptographically, and prove every decision independently.
Principles
The constraints that shape the platform.
Verify before execution
If Intended cannot verify the action, the action does not execute. Verification is not advisory.
Authority must be provable
Every approved action receives a cryptographic Authority Token. Every decision is appended to an immutable audit chain.
Intent has to be legible
Autonomous systems are only governable if what they are trying to do can be expressed, interpreted, and verified in a shared language.
Open source
Core standards and primitives stay open.
The shared language and verification primitives are designed to spread. The enterprise platform is what operationalizes them.
Open Intent Layer
@intended/open-intent-layer
The open standard for defining what agent actions are intended to do across 14 domains and 306 intents.
Apache 2.0
Verification SDK
@intended/verify
Offline verification tooling for Authority Tokens and evidence-chain artifacts.
Apache 2.0
Connector SDK
@intended/connector-sdk
Build custom connectors that bring external systems into the same verification and authority model.
Apache 2.0
Join us in defining this category correctly.
Whether you are building agents, governing them, or auditing them, we want the category language and the system guarantees to be precise.