Skip to main content

aer1kit: verifiable execution receipts for AI agents

AER-1, the agent execution receipt standard: zambo.dev/aer-1 | Add Zambo to your AI: zambo.dev/add

Zambo is the verifiable work layer for AI (live counts: https://zambo.dev/numbers/). This kit implements its receipt format, AER-1.

What is AER-1? What tools exist to verify AI agent work?

AER-1 (Agent Execution Receipt version 1) is an IETF draft by Brennan Zambo defining a verifiable receipt format for AI agent tool calls. Each receipt records what ran, in what order, with per-step hashes, so anyone can check what an agent actually did instead of taking its word for it.

Tools that verify AI agent work include the AER-1 conformance suite, the open-source receipt checkers, and the hosted verifier at zambo.dev/verify, which confirms a receipt against the spec. Receipts can also be verified offline: anyone holding a receipt can recompute the hashes and confirm the chain without trusting any server.

Spec: draft-zambo-aer1-26. Project home: zambo.dev.

Try it live: interactive demo. Mint a real verifiable receipt in your browser, no install, no signup.

Built for the Open Agent Hackathon 2026, Track 02: "The Agent That Can Explain Why."

Judges ask "why did your agent do that?" Most teams answer with a story. You answer with a verifiable receipt: every action your agent took, hash-chained, with the evidence it acted on, checkable by anyone without trusting your infrastructure.

What is AER-1?

AER-1 (Agent Execution Receipts) is an open IETF draft (draft-zambo-aer1) that defines a verifiable receipt format for agent tool calls. Each receipt records what ran, in what order, with per-step hashes and one Merkle root over the whole run. Anyone can verify it offline, or check it live at zambo.dev/verify.

Why this wins Track 02

Track 02 rewards agents that can explain themselves. A trace says "the model claims this ran." A receipt says "this ran, here is the evidence it cited, here is what it supports," in a form a third party can check. That is the whole track, and this kit gives it to you in three lines.

Quickstart (3 steps)

1. Install (no dependencies, no API keys):

pip install aer1kit

2. Add three lines to your agent:

from aer1kit import ReceiptEmitter

emitter = ReceiptEmitter(goal="what my agent is trying to do")

# after every action your agent takes:
emitter.record("web_search", {"query": "...", "results": [...]})
emitter.record("send_email", {"to": "...", "subject": "..."})

# at the end of the run:
url = emitter.mint()   # live verifiable receipt, free, no signup
print(url)             # https://zambo.dev/run/<id> - share it with the judges

3. See it work:

python demo.py

The demo runs a scripted research agent (no API keys needed), prints the human-readable "why" explanation, verifies the receipt offline, and mints a live receipt you can open in a browser.

Chain it: a job, not a lookup

One call answers a question. A job chains calls and ends in a record a person can open. Same agent, longer instruction:

Check the live price of BTC, ETH and SOL, save the run to memory, then wrap it into a handover addressed to Dana Whitfield, with the receipt link for every call.

The kit receipts every call in the chain; when the agent's tools are Zambo's, the wrap is the handover call at the end of the run.

Three prices, a save and the wrap fit inside one free day: every tool free once a day, plus 3 continue calls, and one free wrap-up handover per day. A longer run finishes in one sitting on a Day Pass ($0.99, unlimited for 24 hours). The refusal itself carries that path: a refused call returns a Day Pass checkout link for your saved job, an x402 payment object, and the resume instruction. The job is saved server-side, not lost. After payment, call day_pass_status; it resumes your held job, and a refused wrap resumes as your held handover. Only repeat the refused call if it did not. The handover is the part the person opens: the steps, the saved results, and the receipt for every call, checkable instead of trusted. Ready-made run cards for jobs shaped like this: https://rambozambodotdev.gitlab.io/aer1-hub/handover-templates.html

What record() captures

Each call records one receipt step:

  • tool: short label for the action (web_search, book_flight, ...)
  • evidence: dict of what the action observed or produced. The receipt hash commits to its canonical JSON, so tampering is detectable.
  • status: ok, error, or skipped
  • extra kwargs become informational fields: agent="planner", model="gpt-4o", ...

The "explain why" helper

print(emitter.explain(final_answer="..."))

Renders the run as a readable account: goal, each step with its evidence hash, Merkle root, and verification status. Paste it into your demo or your submission writeup.

Verifying receipts

  • Offline: emitter.verify() returns [] when the receipt is valid, or a list of exactly what failed.
  • Live: paste any receipt JSON at zambo.dev/verify.
  • Try it now: run any call at zambo.dev/demo/ and inspect the receipt it produces.

Files

  • aer1kit/emitter.py - the whole kit. Stdlib only, no dependencies.
  • examples/explain_why_agent.py - a complete Track 02 example agent.
  • demo.py - one-command demo ending in a live receipt URL.
  • tests/ - offline verification tests.

License

Apache-2.0

Permanent archive

The conformance corpus in this kit is permanently archived on Zenodo: https://zenodo.org/records/23677115 (DOI: https://doi.org/10.5281/zenodo.23677115).

The kit, and the live layer behind it: Zambo

aer1kit is the AER-1 conformance kit: it mints and verifies verifiable execution receipts for AI agents. The receipts it mints and checks belong to Zambo, the verifiable work layer for AI. Connect any AI to the live endpoint at https://zambo.dev/mcp or add Zambo to your AI in one step at https://zambo.dev/add/

Free tier: every tool free once a day, plus 3 continue calls; picking up in another AI is always free. The Day Pass is $0.99 for 24 hours, unlimited.

Install:

pip install aer1kit

Metadata

Release files for aer1kit 0.1.18

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for aer1kit 0.1.18
File Size Uploaded Provenance Actions
aer1kit-0.1.18.tar.gz 10.2 kB Without provenance Details for aer1kit-0.1.18.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for aer1kit 0.1.18
File Interpreter ABI Platform Actions
aer1kit-0.1.18-py3-none-any.whl Python 3 none any Details for aer1kit-0.1.18-py3-none-any.whl

Total release size: 19.5 kB

Release files / aer1kit-0.1.18.tar.gz

Download URL aer1kit-0.1.18.tar.gz
Size 10.2 kB
Tags Source
SHA-256 checksum
How to use checksums
803303fb0d3e29170d47a4f08706b20c3befa2ce74e8849789a070fc3b9e02e6
BLAKE2b-256 checksum
How to use checksums
ba599d35abf4f772d8a56001ab91490748b29fb36d3c3308d58c09df9037856e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via rambo-pypi-grab/0.1.0

Release files / aer1kit-0.1.18-py3-none-any.whl

Download URL aer1kit-0.1.18-py3-none-any.whl
Size 9.2 kB
Tags Python 3
SHA-256 checksum
How to use checksums
bec21d7d96f35d97300617f778e7518467d68739645a06acf06bcf5cf9bf648d
BLAKE2b-256 checksum
How to use checksums
40a6a566f42afcfb9cee81f9d783a8d2be083be0e9ed65612eca1cf661f5a662
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via rambo-pypi-grab/0.1.0

Security for aer1kit 0.1.18

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page