aetherproof 0.2.0

Cryptographic, tamper-evident execution receipts for any AI model

AetherProof

Cryptographic, tamper-evident execution receipts for any AI model.

One call generates a 128-byte Ed25519-signed receipt that proves what model ran and what it produced — verifiable offline by anyone, with no server and no account. Works with Claude, GPT, Gemini, or a local model via Ollama.

Install

pip install aetherproof

Drop-in: one line to a verified receipt

Run a local model and get its answer plus a signed, tamper-evident receipt — no extra dependencies, no server:

import aetherproof

r = aetherproof.ollama("llama3", "Explain quantum tunneling")
print(r.text)        # the model's answer
print(r.verify())    # True — cryptographically receipted
r.save("call.receipt")

Cloud providers work the same way (pip install openai / anthropic):

r = aetherproof.openai("gpt-4o", "Summarise this contract.")
r = aetherproof.anthropic("claude-3-opus-20240229", "Explain GDPR Article 13")

Already have a response? Receipt it directly:

receipt = aetherproof.for_anthropic(prompt, response_text, model="claude-3-opus-20240229")
print(receipt.verify())   # True
receipt.save("run.receipt")

What a receipt proves

• What ran — a hash identifying the model.
• What it produced — a hash of the output; any change is detectable.
• Tamper-evidence — flip a single byte and verification fails.
• Offline-verifiable — anyone can check it with the public key; no server.
• Portable — the same 128-byte wire format verifies in Python, the Rust CLI, and the AetherOS kernel.

The receipt can also ride invisibly inside the model's own output text (aetherproof.embed), so the proof travels with the content.

License

AGPL-3.0-or-later. See LICENSE.