Getting Started
Kanoniv has three layers. Pick the path that matches what you are building right now.
Path 1: I'm building an MCP server and need auth
MCP servers currently accept tool calls from any agent with no identity, no delegation chain, and no audit trail. Kanoniv adds cryptographic auth in 5 lines.
- Agent Identity (5 min) - Generate a keypair, get a DID, sign and verify a message.
- MCP Server Auth (5 min) - Add proof verification to your MCP server. TypeScript-focused since most MCP servers are TypeScript.
After this, every tool call includes a cryptographic proof of who the agent is and what it is allowed to do.
Path 2: I'm building agents that delegate to each other
Multi-agent systems need a way for one agent to grant another agent a subset of its authority - with constraints that can only narrow, never widen.
- Agent Identity (5 min) - Generate keypairs and DIDs for your agents.
- Delegation (5 min) - Create delegations with caveats, build chains, and verify them. Covers action scopes, expiry, cost limits, and resource patterns.
After this, your agents carry verifiable authority. Any system can verify the full chain offline - no server calls needed.
Path 3: I need identity resolution
Multiple agents or systems create records for the same real-world entity. Kanoniv resolves them to a single canonical ID - deterministically, with confidence scores and full auditability.
- API Quickstart (5 min) - Resolve your first entity with one API call.
After this, every agent calls POST /v1/resolve/realtime before writing to any system. Same person, same entity ID, every time.
Free sandbox
Create a sandbox account at kanoniv.com/signup - 1,000 entities, full API access, no credit card required.
What connects them
All three paths share the same cryptographic foundation: Ed25519 keypairs, did:agent: decentralized identifiers, and signed message envelopes. An agent that has an identity (Path 1 or 2) can use that same identity to authenticate resolve calls (Path 3). The layers compose.
| Layer | Primitive | Purpose |
|---|---|---|
| 0 - Identity | did:agent: DID | Who is this agent? |
| 1 - Delegation | Caveated authority chains | What is this agent allowed to do? |
| 2 - Resolution | Canonical entity IDs | Which real-world entity is this? |
| 3 - Memory | Cross-agent knowledge | What does this agent know? |
| 4 - Audit | Signed provenance DAG | What did this agent do, and who authorized it? |