These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
Mandare
Mandare is an open-source accountability system for fleets of AI agentsm, which enables users to oversee, constrain, and audit agent actions. It provides mechanisms for managing permissions, spending, and accountability, helping ensure that agent behavior remains aligned with user-defined rules and oversight requirements.
Mandare combines several components. A Passport gives each agent a signed identity. A Mandate sets the agent's authority in machine-readable form, including spending caps, time windows, permitted actions and requirements for human approval. Every request passes through a gateway, which checks the kill switch and the mandate before acting. Each request is recorded in the Ledger before it runs, and its true cost is recorded afterwards. Refused actions are recorded too.
The Ledger is append-only and cryptographically chained, so any edit or deletion is detectable. An external witness, ideally run on separate infrastructure, keeps a record of the ledger's history. It can also anchor this history publicly, so not even the operator can truncate or rewrite records unnoticed. The system fails closed: without a mandate, nothing can be spent, and if the ledger is unavailable, no action runs. Third parties can verify integrity certificates without accessing the ledger. Mandare is written mainly in TypeScript, with a Python client.
About the tool
You can click on the links to see the associated tools
Tool type(s):
Purpose(s):
Lifecycle stage(s):
Type of approach:
Maturity:
Usage rights:
Target groups:
Target users:
Stakeholder group:
Geographical scope:
People involved:
Required skills:
Technology platforms:
Programming languages:
Use Cases
Would you like to submit a use case for this tool?
If you have used this tool, we would love to know more about your experience.
Add use case
Partnership on AI
