- Category
- Cybersecurity
- Rank
- No. 1557Tools index
Previous survey · No. 1583 ·
- Type
- TOOL
- Use case
- Security & Identity
- GitHub
- 15 stars
- Latest release
- v4.1.1
- Date
About
Tkeeper implements a "cryptographic identity" model for AI agents, pairing a private key with a manifest that defines exactly which actions it may sign; an agent must request a signature for a specific action and receives it only after policy checks pass. It uses threshold cryptography to split key control across multiple nodes so no single compromised node can steal the key or alter policies, with pluggable modules for agentic payments (Google AP2, Mastercard Verifiable Intent) and crypto wallet actions.
At a glance
TKeeper applies policy and approval checks before signing actions for agents or services. It is worth evaluating if your application can verify a signature before executing a sensitive command.
Good fit
Teams designing a service that can separate action approval from execution.
Consider something else when
Use a different integration if the receiver cannot bind the signature to the exact action and enforce replay protection.
Start here
Use a test key and a synthetic command. Check one approved request, one rejected request and one replay at the receiving service.
Cost and requirements
Self-hosted signing infrastructure with documented single-node and quorum modes. It needs an application integration that verifies signed actions.
Limits to know
The signing service returns a signature or signed payload. The receiver still has to validate freshness, action contents and replay rules before execution.
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.