SmartPolicyTranslator

Translate natural-language regulation into REAL, schema-conformant IAIso policy files — and wire them straight into your architecture.

Regulation is prose. IAIso deployments run on policy files. SmartPolicyTranslator closes that gap: it ingests regulatory text, parses each clause's deontic force (MUST / SHOULD / MAY), and emits a valid IAIso v1 policy document — with full provenance from every policy choice back to its source clause.

 ingest ─▶ parse (deontic) ─▶ translate ─▶ validate ─▶ audit
  text/PDF/URL   MUST/SHOULD/MAY   → IAIso policy   vs real schema   provenance

This is not an invented format. The IAIso policy schema, normative spec, and official test vectors are vendored under spec/iaiso/ directly from github.com/SmartTasksOrg/IAISO (IAIso-v5.0). Every emitted policy is validated against them, and the loader passes all of IAIso's own policy test vectors.

Install

SmartPolicyTranslator is not published on PyPI or any other package registry yet. Until this section says otherwise, a package called sf-smartpolicytranslator on any registry is not ours, and neither is smartpolicytranslator.

Install from a clone (Python 3.10 or later):

git clone https://github.com/SmartTasksOrg/sf-smartpolicytranslator
cd sf-smartpolicytranslator
python -m venv .venv
. .venv/bin/activate          # Windows PowerShell: .\.venv\Scripts\Activate.ps1
python -m pip install .
python -m sf_smartpolicytranslator.cli examples/sample_regulation.txt

Status

• Version 1.0.0, experimental. Translates regulation text into IAIso v1 policy files that are validated against the IAIso schema and test vectors vendored in spec/iaiso/; 4 tests.
• Published: nowhere yet; install from a clone (above).
• Tested: the 4 tests in tests/ on Python 3.12, Linux, on every push to main and every pull request (.github/workflows/ci.yml).
• Not tested: Windows and macOS; the LLM-assisted parsing path; Python versions other than 3.12.
• Ports: The client ports in ports/ have no automated check; none is published on a registry.
• Security review: none independent. Report vulnerabilities as described in SECURITY.md.

Ready to run

The core needs zero third-party packages (stdlib only):

export PYTHONPATH=src
python3 -m sf_smartpolicytranslator.cli examples/sample_regulation.txt
# → a valid IAIso policy: version "1", enforcement_mode, consent scopes, pressure…

# LLM-assisted parsing (LM Studio / Ollama / llama.cpp):
python3 -m sf_smartpolicytranslator.cli --llm --provider lmstudio examples/sample_regulation.txt

# REST service:
bash scripts/run.sh          # → http://localhost:8000/docs   (or: docker build/run)

What the translation produces

A regulation with MUST/high-risk obligations yields a strict, fail-closed IAIso policy: enforcement_mode: "strict", consent.required_scopes derived from the data/access obligations (in IAIso's scope grammar, e.g. data.pii.read, model.invoke, decision.override), a tightened pressure posture (escalation_threshold lowered, post_release_lock on), a shorter consent TTL, and a metadata.provenance[] trail mapping each choice to its clause. A permissive regulation yields a permissive policy.

Integrations & ports

SmartPolicyTranslator produces IAIso policy; IAIso enforces it.

• LangChain (+ real IAIso SDK) — integrations/langchain/: bounded_execution_from_regulation() translates regulation and loads the policy into IAIso's real BoundedExecution + IAIsoCallbackHandler, so a LangChain LLM is governed by the translated policy (pip install iaiso[langchain]).
• n8n / Flowise — integrations/n8n/, integrations/flowise/.
• Language ports — ports/{node,react,java,go,php,ruby,rust,csharp,swift,c,cpp}, mirroring every language IAIso itself ships (plus React and C/C++) so a policy can be generated and enforced in the language of your IAIso deployment. See ports/README.md for the port↔iaiso-* map.
• Local LLMs — src/sf_smartpolicytranslator/llm.py: one OpenAI-compatible client for LM Studio, Ollama, and llama.cpp.

Tests

python3 -m pytest tests/ -q     # translation + REAL IAIso schema/vector conformance

These run on the standard library alone — no iaiso package or network required.

Who's behind this

• Roen Branham — CEO & AI Strategy Architect · CISSP-certified AI, security & governance architect; author of IAIso and sole inventor of the Z4 Semantic Fabric patent application. LinkedIn
• Le Vu Tanh — CTO & Core Engineering Lead · Chief architect of the Cortex engine; large-scale system reliability and low-latency infrastructure — the engineer who ships what gets architected. LinkedIn

The team behind IAIso & SmartTasks: a CISSP-certified security & governance architect and a large-scale systems engineer — 20+ years shipping secure, AI-driven platforms for regulated, blue-chip environments (Allianz, BMW, Rolls-Royce, Heidenhain).

Runs on governed local models

Structured, machine-readable scorecards make model choice precise instead of guesswork.

This tool is local-first, so pair it with models you can actually vet. SmartTasks publishes 21+ governance-validated GGUF builds on Hugging Face — each with a machine-readable scorecard (capability tiers L1 Layman → L5 Agentic, IAIso conformance invariants (pass/warn/fail), OWASP-mapped garak red-team, transparency probes (viewpoint-alignment / over-refusal), and per-file SHA-256). Gate model selection on evidence, not vibes — and every finding, including warnings, is published in full.

→ SmartTasks on Hugging Face · Qwen3.6-27B (L5 agentic) · react-agent-coder-llama-3.1-8b (agentic coder) · gpt-oss-20b (open reasoning)

Get in touch

• Companies & enterprises: enterprise@smarttasks.cloud — we help teams integrate SmartPrompt + IAIso into their architecture so governance and audit-readiness become a byproduct of how they already work.
• The standard: IAIso · iaiso.org
• The product: SmartTasks.cloud

Built by SmartTasks Lab. Apache-2.0. Contributions welcome.