AI that can't lie,
leak, or go rogue
O-Lang was forged in the world's most demanding deployment conditions — multilingual populations, fragmented regulation, zero-tolerance compliance. A governed runtime that enforces what your AI can do, automatically redacts PII, and produces a signed audit trail for every execution. If it works here, it works anywhere.
Already using LangChain, CrewAI, or OpenAI Agents? O-Lang governs their tool calls without replacing them.
// O-Lang — Bank Balance Workflow Workflow "Bank Balance Check" with user_question, customer_id, locale Allow resolvers: - @o-lang/bank-account-lookup - @o-lang/currency-format - @o-lang/llm-groq Step 1: Ask @o-lang/bank-account-lookup "{customer_id}" Save as account_info Step 2: Ask @o-lang/currency-format "{account_info.balance}" "{locale}" Save as formatted_balance Step 3: Ask @o-lang/llm-groq "{user_question}. Balance is {formatted_balance}." Save as response Return response
Workflow domains in v1.1
Cryptographic audit coverage
Hallucinations in governed scope
Open source license
Write it once.
The kernel enforces it forever.
Declare your workflow in O-Lang. The kernel enforces every tool call, intent boundary, and safety constraint at runtime — deterministically, with a cryptographic receipt. No framework needed.
Wrap your existing tools
Write a .ol file. Declare your inputs, the tools it may call, and the intent scope. The kernel reads this as the authoritative contract — nothing outside it runs.
Kernel intercepts every call
The kernel sits between your orchestration layer and your tools. Every call is validated against the allowlist, intent-checked, and safety-filtered — before it reaches the tool. Your framework never knows.
Cryptographic audit trail
Every execution is SHA3-256 fingerprinted and ED25519-signed by the kernel. Your outputs are verifiable, tamper-evident, and ready for compliance.
Everything a governed
AI runtime needs
Tool Allowlists
Every workflow declares exactly which tools it may call. Any unlisted tool is rejected by the kernel at runtime — no exceptions.
@governance.allowed_toolsVerified Intent
Declare scope and prohibited actions. The kernel enforces both before and after execution — zero tolerance for scope creep.
__verified_intentCryptographic Audit
Every execution produces a SHA3-256 fingerprint, chained to the previous, and signed with ED25519. Full deterministic proof.
sha3-256 · ed25519Symbol Resolution
Steps compose with typed symbols. Each tool declares its output schema — downstream steps receive validated, typed data, never raw LLM output.
$step.symbol.fieldPII Detection & Safety
The kernel detects and redacts PII automatically on every execution — no extra tool needed. Covers 20+ languages including Yoruba, Igbo, Hausa, Swahili, and Zulu. Hallucinations and prompt injection blocked at the same layer.
kernel-native · 20+ languagesTool Registry
Publish, version, and discover tools at olang.cloud/registry. Every tool is signed and schema-validated before listing.
@o-lang/tool-nameWhere O-Lang Works
Today
Governed AI workflows for critical domains. Built for the world's most demanding deployment conditions — multilingual populations, fragmented regulation, zero-tolerance compliance.
Complex financial systems, 500+ languages, fragmented regulation, and high-stakes compliance — these conditions forced us to build governance features that every enterprise globally needs but hasn't demanded yet. Africa surfaced the problem. O-Lang solves it for everyone.
Financial Services
Safe customer assistants that answer balance questions, explain policies, and block unauthorized transfers — with locale-aware formatting across any currency and language, out of the box.
Same workflow · any locale
Healthcare & ICU
Patient admission workflows that enforce team scopes, prevent unauthorized actions, and log every decision cryptographically — built for compliance whether you're in Lagos, London, or Los Angeles.
Workflow
- • Verify team authorization for ward access
- • Check bed availability in real-time
- • Summarize admission for clinical handoff
- • Block treatment suggestions outside scope
- • Cryptographic log for every decision
Government & IoT
Audit-ready workflows for public services: permit approvals, sensor data validation, and citizen queries — with cryptographic proof of every action, across any language or jurisdiction.
Use Case
- • Validate citizen ID before service access
- • Cross-reference sensor data with policy rules
- • Generate tamper-evident audit logs
- • Escalate exceptions to human reviewers
- • GDPR, NDPR & ISO 42001 aligned by default
Built for complexity · Deployable everywhere
Add a governed chatbot to any website in 3 minutes. No framework required. Works on WordPress, React, static HTML — or integrate deeper via the SDK.
Governance that speaks
your language
O-Lang's kernel detects and redacts PII automatically on every execution — in 20+ languages, with first-class support for African languages that every major AI framework ignores. No extra configuration. No optional plugin.
African Languages
First-class support — not an afterthought
Yoruba
West Africa
Igbo
West Africa
Hausa
West Africa
Swahili
East Africa
Zulu
Southern Africa
Xhosa
Southern Africa
Amharic
East Africa
Somali
East Africa
Detects names, phone numbers, national IDs, addresses, and financial identifiers written natively in each language — not transliterated, not approximated.
Also supported
workflow · kernel auto-redacts PII
Workflow "Customer Support" with message, lang
// No PII resolver needed — kernel handles it// automatically for all 20+ supported languages
Allow resolvers:- @o-lang/llm-groq
Constraint: lang in ["yo", "ha", "sw", "ig", "en"]
Step 1: Ask @o-lang/llm-groq "{message}" ↑ kernel redacted PII before this ranSave as response
Return response20+
Languages supported
including 8 African languages
8
African languages — live
Yoruba, Igbo, Hausa, Swahili, Zulu, Xhosa, Amharic, Somali
1.4B+
People represented
across supported African languages
Write and run
.ol files in your
browser
No install. No config. Connect to the O-Lang cloud kernel and execute governed workflows live.
- Monaco editor with O-Lang syntax highlighting
- Real-time execution trace — see every resolver step
- Live safety logs and cryptographic audit output
- Save and share workflows with a permanent URL
// ICU Admission — Healthcare workflow workflow icu_admission_v1 { @governance { allowed_resolvers: [ "@o-lang/icu-admission", "@o-lang/llm-groq" ] scope: "team-a-icu-only" prohibited_actions: [ "unauthorized_procedure", "ward_transfer_without_auth" ] } step admit = @o-lang/icu-admission({ patient_id: $input.patient_id, ward: $input.ward, team: $input.team }) }
O-Lang vs Alternatives
LangChain builds pipelines.
LlamaIndex builds indexes.
O-Lang governs both.
O-Lang wraps your existing AI stack with governance, cryptographic audit trails, and provider portability — allowing teams to switch AI models without disrupting workflows or rebuilding infrastructure.
Provider Portability
AI providers change.
Your workflows shouldn't.
OpenAI today. Claude tomorrow. Local models next year. O-Lang abstracts execution so your applications remain stable even when models, APIs, or providers evolve.
What it does
Builds LLM pipelines and chains tool calls together
Governance gap
No enforcement of what tools can actually do. Any step can call anything.
O-Lang adds
Governs LangChain's tool calls, enforces resolver allowlists at runtime, and lets teams swap AI models without rewriting workflows.
What it does
Indexes documents and retrieves context for RAG
Governance gap
No policy layer. Retrieved context flows into LLM without validation.
O-Lang adds
Wraps LlamaIndex retrieval with audit trails, output safety checks, and provider portability across models.
What it does
Orchestrates multi-step AI agents with tool use
Governance gap
No cryptographic proof of what ran. No replay. No tamper-evident log.
O-Lang adds
Adds replayable execution, Merkle-signed audit trails, and stable workflows even when switching AI providers.
What it does
Validates LLM outputs against schemas and rules
Governance gap
Output-only. Doesn't govern the workflow or resolver chain that produced it.
O-Lang adds
Governs the full execution lifecycle — inputs, resolvers, outputs, and model portability across providers.
Feature
LangChain
LlamaIndex
OpenAI
Guardrails
O-Lang
Resolver allowlists
Cryptographic audit trail
Replayable execution
Merkle + RSA signing
Swap models without rewrites
Provider portability
PII redaction 20+ langs
Works with existing tools
Multi-locale formatting
ISO 42001 aligned
Governance score
O-Lang is not a replacement — it's a portable AI governance layer. Use it alongside any framework, model, or provider you already run.
Every tool is signed,
versioned, and governed
O-Lang has its own tool ecosystem. Each tool (called a resolver internally) is a versioned, schema-validated package published to the O-Lang registry — not a wrapper around someone else's SDK.
o-lang kernel — your execution layer
underlying APIs (tools call these)
your tools — completely unchanged
Your frameworks call tools. O-Lang intercepts those calls, enforces governance, and returns signed output. Nothing in your existing stack changes.
@o-lang/llm-groq
Governed LLM calls with intent enforcement and output validation
@o-lang/bank-account-lookup
Read-only financial data tool with audit-signed output
@o-lang/icu-admission
Clinical workflow tool with team-scope enforcement
@o-lang/semantic-search
Vector retrieval with schema-validated, typed output
@o-lang/legal-extractor
Contract clause extraction with prohibition enforcement
your-tool
Publish any tool to the registry — HTTP, any language
JSON in, JSON out. That's the whole contract.
If your tool speaks JSON, O-Lang can govern it. Python, Go, Rust, Java, Bash, WASM — the kernel is indifferent to language. It only cares about the shape of the data, not who produced it.
Discover and publish
governed tools
Every tool in the registry is schema-validated, version-pinned, and signed. Drop it into any workflow with a single import.
@o-lang/bank-account-lookup
Secure account balance and transaction lookup with PII validation and multi-jurisdiction audit compliance.
@o-lang/icu-admission
ICU bed management, team-scoped admission, and HIPAA-compliant patient data handling.
@o-lang/llm-groq
Fast, governed LLM inference via Groq. Output schema-validated and hallucination-filtered.
@o-lang/notify-telegram
Send governed notifications to Telegram channels or users. Rate-limited, audit-logged, and allowlist-enforced.
@o-lang/semantic-doc-search
Vector-based semantic document search with scoped retrieval. Prevents out-of-context document access.
@o-lang/js-olang-tester
Unit and integration test harness for O-Lang workflows. Simulates resolver chains and validates governance policy enforcement.
@o-lang/python-olang-tester
Python test harness for O-Lang workflows. Available on PyPI. Validates resolver behaviour and policy compliance from Python environments.
@o-lang/legal-extract
Extract-only legal document analysis. Blocks legal advice output under GDPR, UK Companies Act, US Federal, and other jurisdictions.
@o-lang/llm-openai
OpenAI GPT-4o resolver with output schema enforcement and rate-limit-aware execution.
@o-lang/pii-validator
Global PII validation across 40+ identity formats: SSN, NIN, BVN, NHI, TFN, Aadhaar, and more.
@o-lang/device-commander
Govern device command dispatch to IoT endpoints — validates command signatures, enforces actuation allowlists, and logs every instruction to the audit chain.
@o-lang/sensor-router
Route sensor data events through governed alert pipelines — threshold checking, anomaly detection, and conditional escalation with full traceability.
@o-lang/infra-monitor
Infrastructure health monitoring — aggregates uptime, latency, and error rate signals into governed status reports with SLA enforcement.
Governed AI for
high-stakes domains
O-Lang was designed for industries where AI errors are not just inconvenient — they are dangerous, illegal, or catastrophic. Finance. Healthcare. Government. IoT.
Finance
Balance inquiries, transaction classification, fraud detection — with multilingual PII redaction (including Yoruba, Hausa, Swahili), hallucination prevention, and audit-ready compliance trails.
Healthcare
ICU admissions, clinical triage, ward management — with team-scoped access, HIPAA filters, and zero-tolerance prohibited procedure enforcement.
Legal
Contract extraction, NDA review, clause identification — extract-only mode blocks legal advice output, protecting against unauthorized practice of law.
IoT & Infrastructure
Device command governance, sensor data pipelines, automated alert routing — with strict tool allowlists preventing unauthorised actuation and real-time audit trails for every device action.
Everything you need
to build governed AI
From installing the kernel to publishing your first tool — step-by-step guides, API reference, and real-world examples.
Your AI should be
as accountable as you are
Join researchers, engineers, and enterprises building the next generation of governed AI systems on O-Lang.
Get in touch
Interested in research collaborations, enterprise pilots, or integrating O-Lang into your platform?
info@olang.cloud9
countries
20+
languages
MIT
licensed
∞
audit trail
Built in public.
Used in production.
Stay ahead of
governed AI.
Get notified when we ship new resolvers, compliance frameworks, and deployment guides — plus AI training updates and developer onboarding tutorials as they drop. No spam — just signal.
No spam. Unsubscribe any time.
Join the community
Built in public from day one. Every resolver, every kernel decision, every compliance framework — open source, auditable, and MIT licensed. Africa surfaced the problems. The solutions belong to everyone.