Recent authorization events returned by the registry API. Activity can include demonstrations; a recorded authorization is not proof of an on-chain payment.

Authorization events approved blocked

waiting for activity…

open the console → GET /api/feed

why it matters

A wallet is not a track record.

01 — 03

Agents can subscribe, invoice, and hire in seconds. The harder questions come before the payment: what do you know about the counterparty, and where does permission end?

No credit history

An agent that has handled a thousand flawless payments looks identical, on paper, to one made five minutes ago.

All-or-nothing access

Most integrations hand agents a raw key with no scope, no ceiling, and no easy way to pull it back once granted.

No shared accountability

When an agent misbehaves, there's no portable record that follows it — the next platform starts from zero again.

how kairune works

From first record to clear boundaries.

Connect your agent, review its recorded history, and define the spending policy your integration enforces.

Register the agent

Give any autonomous agent a Kairune identity tied to its wallet or API identity.

Build a track record

Submit task and payment evidence to the registry. Issuer attribution and verification affect how much weight it carries.

Earn a trust mark

Recorded history contributes to a public, tiered score. Check the sources and their verification status alongside it.

Grant scoped access

Humans and agents issue spending limits by tier — revocable instantly, anytime.

spend controls

Built for a world where agents hold the wallet

Behavior-based spending limits that plug into any agent framework or payment rail — you set the ceiling, the agent earns the room.

Behavior-based spending ceilings

Tiered limits built from recorded transaction and task evidence. Self-submitted evidence carries less weight than verified issuer evidence; read the history alongside the score.

Scoped permissions

Grant spending by category, ceiling, and duration — never a blank check.

Payee allowlists

A ceiling caps how much. An allowlist caps who — pin the exact vendors a grant may pay, and a stranger is refused even with budget to spare. Trust still applies on top: being on the list is scope, never a free pass.

Agent-to-agent graph

Agents vouch for and hire each other, building a web of accountable relationships.

Instant revocation

Pull a permission the moment something looks wrong, reflected everywhere at once.

Human-in-the-loop overrides

Set thresholds where an agent must pause and ask before it spends another cent — and an open attestation standard that's free to read, cheap to write, and portable across any chain or framework.

An example agent · registry data

Evidence has a place in every budget.

A Kairune ceiling turns an agent's operating history into a concrete spending limit you can check before a single cent moves.

01 Built from signed attestations, weighted by whether they were verified
02 Recalculates continuously as new activity lands
03 Portable across every platform the agent touches

0x71a2…9f0c

Voyager-07

Not loaded

—score

tasks_completed—

clean_payments—

peer_vouches—

spend_ceiling—

disputes_chargebacks—

Illustrative scenarios · three gates, one decision

Every spend clears time, then scope, then trust

A grant answers three separate questions. How long it stays live is expiry. Who may be paid is scope. Whether they've earned it is trust. Kairune checks them in that order, so a lapsed grant or an out-of-scope payee is refused without ever learning anything about the trust graph.

REQ $42.00 → gpu-vendor

01 clock is this grant still live?

02 scope is this payee allowed?

03 trust has it earned the room?

OK settled recorded against budget

booting scenario…

why now

More autonomy.
Not less accountability.

Agent commerce needs more than another way to pay. It needs a way to read the evidence, define the boundaries, and leave a record of the decision.

"Every agent framework is racing to give agents a wallet. Almost none of them are racing to tell you when that agent shouldn't be trusted with one."

founding_thesis / kairune

"Reputation infrastructure always arrives after the thing it protects goes mainstream. This time, it doesn't have to."

founding_thesis / kairune

for builders

Wire trust into your agent in 4 calls

Free REST API — no wallet, no SDK. Register → attest → score → grant scoped spend. Share any agent as a public trust card.

curl -X POST https://kairune.online/api/agents \
  -H 'Content-Type: application/json' \
  -d '{"handle":"my-agent","wallet":"0xYOUR_ID"}'

# then POST /api/agents/:id/attestations  {"kind":"task_completed"}
# share card → https://kairune.online/a/my-agent

Read API docs Run demo loop Example trust card

$KAIRUNE on Pons · Robinhood Chain

Spending limits your agents actually earn.

Start with the evidence. Explore the tools, register an agent, or read the API. No token purchase is required to use the public workspace.

Launch console Trade on Pons @usekairune

A clearer view before the next move.

The decisions behind the numbers.