AltmanAI SDK

The official developer toolkit for building human-first AI apps with verifiable impact.

AltmanAI SDK gives developers a trusted integration layer for building AI products that are transparent, explainable, and measurable. It provides shared schemas, receipt generation, artifact verification, AINet event hooks, and explainable execution patterns across TypeScript and Python.

Built for a future where humanity leads and intelligence follows.

This is also the canonical home for AltmanAI's verifiable-artifact tooling — receipt.py and verify.py here provide deterministic SHA-256 hashing, receipt generation, and artifact verification. (The formerly separate PAIHI-Portables repo covered this same ground with no working code; it's now archived in favor of this one.)

How this maps to P.A.I.H.I.

	Dimension	Where it lives here
P	Proof	receipt.py — every action gets a hashed, timestamped, verifiable receipt
A	Alignment	explainability.py + schemas — decisions carry structured reasoning, not just output
I	Integrity	verify.py — deterministic hashing means an artifact can't silently change
H	Humanity	AINet event hooks — every meaningful action is surfaced for human visibility
I	Impact	Cross-language (TS + Python) so any AltmanAI product can adopt this in production, not just as a demo

---

Why AltmanAI SDK exists

Most AI tooling focuses on speed and output.

AltmanAI SDK focuses on trust, visibility, and real-world accountability.

This repository exists to help developers build systems that do more than generate responses. It helps them create applications that can:

• produce verifiable receipts
• validate artifacts and outcomes
• expose structured reasoning metadata
• emit trustworthy event logs
• support human oversight by design

The goal is simple: make it easier to build AI software that people can actually trust.

---

Core capabilities

1. Shared schemas

Standardized schemas for artifacts, receipts, events, and explainability metadata.

2. Receipt generation

Create structured records of important system actions, outputs, and decisions.

3. Artifact verification

Validate whether a file, payload, or output matches an expected structure, fingerprint, or integrity rule.

4. AINet event hooks

Emit consistent events for tracking meaningful activity across AltmanAI systems and integrations.

5. Explainable execution patterns

Wrap outputs with clear metadata such as decision, reason, evidence, timestamp, and actor.

6. Cross-language support

Use the same conceptual model in both TypeScript and Python.

---

Repository layout

altmanai-sdk/
├─ .github/
├─ docs/
├─ examples/
│  ├─ node-basic/
│  └─ python-basic/
├─ packages/
│  ├─ js/
│  │  └─ core/
│  └─ python/
│     └─ altmanai/
├─ schemas/
└─ README.md

---

Quick start

JavaScript / TypeScript

pnpm install
pnpm --filter @altmanai/core build

Python

python -m venv .venv
source .venv/bin/activate
pip install -e packages/python/altmanai
pytest packages/python/altmanai/tests

---

Example use cases

Generate a receipt

Create a structured receipt each time an important AI action occurs.

Examples:

• report generated
• task prioritized
• artifact created
• decision recommended
• verification completed

Verify an artifact

Check whether a document, payload, or generated object conforms to expected schema and integrity rules.

Emit an AINet event

Track meaningful product activity in a consistent format.

Examples:

• artifact.created
• receipt.generated
• verification.passed
• verification.failed
• impact.logged

Add explainability metadata

Attach machine-readable context to an action or output so a developer, auditor, or end user can understand what happened.

---

JavaScript example

import { createReceipt, emitEvent } from "@altmanai/core";

const receipt = createReceipt({
  type: "artifact.created",
  actor: "DailyPilot",
  summary: "Generated daily planning artifact",
  source: "dailypilot-engine",
});

const event = emitEvent({
  name: "receipt.generated",
  source: "dailypilot-engine",
  payload: receipt,
});

console.log(receipt.id, event.id);

---

Python example

from altmanai import create_receipt, emit_event

receipt = create_receipt(
    type="artifact.created",
    actor="DailyPilot",
    summary="Generated daily planning artifact",
    source="dailypilot-engine",
)

event = emit_event(
    name="receipt.generated",
    source="dailypilot-engine",
    payload=receipt,
)

print(receipt.id, event.id)

---

Status

Current status: Initial public foundation

This repository is the official SDK layer for AltmanAI systems. The current version focuses on:

• shared schemas
• receipt generation
• verification utilities
• event hooks
• examples for TypeScript and Python

As the project matures, this repo will become the standard developer integration layer across AltmanAI products and partner workflows.

---

Roadmap

v0.1

• repository foundation
• shared JSON schemas
• core receipt model
• basic verification utilities
• JavaScript package scaffold
• Python package scaffold

v0.2

• AINet event client
• explainability helper models
• richer validation and hashing utilities
• first example apps

v0.3

• framework adapters
• expanded docs
• integration recipes for AltmanAI products
• improved testing and release workflows

---

Security

If you discover a security issue or a vulnerability, please follow the project security policy instead of opening a public issue.

Security contact and disclosure guidance live in SECURITY.md.

---

License

Licensed under the Apache License 2.0.

---

AltmanAI

AltmanAI is building human-first AI systems designed for real-world usefulness, transparency, and measurable impact.

Humanity leads. Intelligence follows.